A new release of Ubuntu Server, version 8.04, is scheduled to be available for free download Thursday with more enterprise qualities than previous releases. The move puts Canonical's Ubuntu, heretofore a consumer-oriented, desktop version of Linux, on more of a collision course with Red Hat Enterprise andNovell Suse Linux Enterprise.
Ubuntu Server 8.04 will be supported for five years, only the second version of Ubuntu to be guaranteed Long Term Support, with the previous 6.06 version also receiving the LTS designation. Ubuntu follows its own naming convention on releases, using the year followed by the month to come up with 6.06, released in June 2006, or 8.04 released in April 2008.
Surprisingly, this version of Ubuntu also appears precertified to run on x86 servers from Sun Microsystems, the first server manufacturer to certify Ubuntu on its hardware. They include the Sun Fire X2100 M2, X2200 M2, and the Sun Fire X4150.
"The release of 8.04 is targeted at the enterprise," said Barton George, group manager of free and open source software at Sun. So the company decided to add it to the same certification process Sun uses with enterprise distributions Red Hat and Suse.
Version 8.04 has been combed for stability issues and tested to make it more reliable, said Mark Shuttleworth, the London-based CEO of Canonical, distributor of Ubuntu. "We commit to bring out a new LTS version every two years. We hope we can establish a cadence with these releases," he said in an interview.
The Sun certification covers its one-, two-, and four-way x86-based server models. Ubuntu hasn't invested in optimizing its system for server operation beyond the four-way models, Shuttleworth said. At the same time, it is talking toDell ( Dell), HP, and IBM about getting them to offer certification that Ubuntu is ready to run out of the box on their x86-based hardware as well.
Ubuntu 8.04 has key Windows integration features. Through a third-party product from Likewise Software, Likewise Enterprise, Ubuntu 8.04 can use services from Microsoft's Active Directory user identification system.
"The key thing is integration with Windows networks," said Shuttleworth. The Likewise product makes it possible to administer Linux through the same management console as Windows administrators are accustomed to using.
It is distributed with the Open JDK or Java Development Kit, which means Ubuntu for the first comes with a Java virtual machine and can run Java applications loaded onto it.
The Ubuntu Server kernel has been hardened against intruders and the 8.04 distribution will include Novell's AppArmour open source policy setting and enforcement software, Shuttleworth said.
In addition, the distribution includes the open source Alfresco content management application, Bacula network backup, Parallels virtualization, Qumranet desktop virtualization based on KVM, Tresys security, Zend Technology's PHP, and Zimbra online e-mail and applications.
When it comes to virtualization, Ubuntu makes use of the kernel with KVM in it; it supports the open source Xen hypervisor as a separate kernel packet. It also has partnered with VMware to ensure VMware's ESX hypervisor runs with Ubuntu 8.04, he said.
"Our approach is not to compete with the virtualization vendors," he noted.
Thursday, April 24, 2008
ProCurve IDM and Microsoft NAP Team Up
ProCurve Networking by HP has announced that the ProCurve Identity Driven Manager (IDM) policy management tool is now interoperable with Microsoft’s Network Access Protection (NAP) technology, providing an integrated, standards-based solution for robust network security. The combined technology solution helps customers to implement customized network access policies for admission of users and clients in wired and wireless networks.
The IDM policy management tool enables network administrators to centrally define and apply policy-based network access rights that allow the network to automatically adapt to the needs of users and devices as they connect. It enforces network security while providing appropriate access to network users and devices. Microsoft NAP is a policy enforcement technology built into Windows Vista and Windows Server 2008 operating systems, and it protects network assets by enforcing compliance with network health policies.
“Having ProCurve switches, access points and IDM complement Microsoft’s NAP architecture provides customers with robust network security,” said Mauricio Sanchez, chief security architect, ProCurve Networking by HP. “Through this integration, ProCurve continues to deliver on its ProActive Defense strategy, which combines access control and network immunity with trusted architecture to defend against potential security threats.”
In addition, ProCurve is the global technology partner for Microsoft Technology Centers (MTCs). MTCs were established in 2001 to showcase Microsoft and interoperable partner technologies that customers can use to plan, design and evaluate customized products and solutions. ProCurve is the sole network infrastructure vendor for all MTCs worldwide. The ProCurve infrastructure in each MTC allows customers to test their products and solutions in a highly available and secure network environment. In addition to ProCurve IDM, the MTCs have deployed ProCurve Gigabit Ethernet switches, access points and the ProCurve Manager Plus network management tool at several locations around the world.
This announcement further strengthens the technology collaboration between ProCurve and Microsoft and reinforces the open standards based solution delivered by ProCurve to meet customer and business requirements.
The IDM policy management tool enables network administrators to centrally define and apply policy-based network access rights that allow the network to automatically adapt to the needs of users and devices as they connect. It enforces network security while providing appropriate access to network users and devices. Microsoft NAP is a policy enforcement technology built into Windows Vista and Windows Server 2008 operating systems, and it protects network assets by enforcing compliance with network health policies.
“Having ProCurve switches, access points and IDM complement Microsoft’s NAP architecture provides customers with robust network security,” said Mauricio Sanchez, chief security architect, ProCurve Networking by HP. “Through this integration, ProCurve continues to deliver on its ProActive Defense strategy, which combines access control and network immunity with trusted architecture to defend against potential security threats.”
In addition, ProCurve is the global technology partner for Microsoft Technology Centers (MTCs). MTCs were established in 2001 to showcase Microsoft and interoperable partner technologies that customers can use to plan, design and evaluate customized products and solutions. ProCurve is the sole network infrastructure vendor for all MTCs worldwide. The ProCurve infrastructure in each MTC allows customers to test their products and solutions in a highly available and secure network environment. In addition to ProCurve IDM, the MTCs have deployed ProCurve Gigabit Ethernet switches, access points and the ProCurve Manager Plus network management tool at several locations around the world.
This announcement further strengthens the technology collaboration between ProCurve and Microsoft and reinforces the open standards based solution delivered by ProCurve to meet customer and business requirements.
Terracotta Offers Alternative to Database Congestion for Java Apps
A small firm, Terracotta, is bringing Java know-how to managing applications in a way that improves their scalability while leaving the application code unchanged. In effect, Terracotta clusters Java Virtual Machines underneath the application, tracks JVM activity, and caches the Java objects and data so that any machine in a server cluster may retrieve them from system memory.
As user traffic to an application builds, this approach has the effect of building a bypass to the congestion center of the typical Java application -- the time-consuming calls to the relational database for data with which to build temporarily needed software objects, and the data they will consume.
Instead of waiting for interactions between the database on disk and application server, Terracotta serves as an intermediary or a kind of alternative database, where data -- particularly scratch data, not the transaction result but all the objects and data needed to execute the transaction -- can be stored in memory until they are needed again or can be safely flushed.
And Terracotta performs these functions not just for one Java Virtual Machine but for as many as needed to meet demand. By clustering the JVM instead of copies of the application, Terracotta simplifies and streamlines many aspects of configuring for application performance. It allows more servers to be added to the machine cluster as traffic builds.
"They offload functionality that the database typically gets bogged down with," said Brad Shimmin, an analyst with Current Analysis. And Terracotta's approach "doesn't require the developer to do anything." Terracotta, on its 2.6 release since Monday, gets added to the application infrastructure, monitors the functioning Java Virtual Machines, and manages their data and object traffic for a server cluster.
"It's not a case of the developer saying, 'Oh, my God, I've got to embed another Java library in my application'," said Shimmin.
Terracotta is doing the opposite of the server consolidation that is going on elsewhere in the data center, said Amit Pandey, Terracotta CEO, in an interview. Instead of subdividing a physical machine by adding applications, Terracotta 2.6 is multiplying JVMs to make one server handle more traffic for one application. As in other forms of clustering, the Terracotta approach allows a JVM to fail without losing its transactions or data, shifting them over to another operative JVM.
Terracotta keeps track of objects and data in server memory and shares that information across a server cluster. Its function in some ways is duplicated by an in-memory database, such as Oracle's TimesTen, that can sit in front of a disk-based system and speed performance. Given its data tracking abilities, it also resembles Oracle Tangosol, which links up server memories into a grid and manages data for the group.
But Terracotta's expertise in clustering Java Virtual Machines springs from the experience of Ari Zilka, a co-founder and chief technology officer of the San Francisco firm, and veteran of scaling up Wal-Mart's Web site operations a few years ago. Unlike Oracle, Terracotta 2.6 is also open source code. It has been deployed by Comcast, Pearson VUE, PartyPoker, and other large-scale Java application users.
Pandey says he knows of 150 deployments, most of which have occurred over the past year as the product became open source code. Terracotta is available in commercially supported versions for $2,000 to $10,000 per node, with deals averaging $7,000 per node, Pandey said.
As user traffic to an application builds, this approach has the effect of building a bypass to the congestion center of the typical Java application -- the time-consuming calls to the relational database for data with which to build temporarily needed software objects, and the data they will consume.
Instead of waiting for interactions between the database on disk and application server, Terracotta serves as an intermediary or a kind of alternative database, where data -- particularly scratch data, not the transaction result but all the objects and data needed to execute the transaction -- can be stored in memory until they are needed again or can be safely flushed.
And Terracotta performs these functions not just for one Java Virtual Machine but for as many as needed to meet demand. By clustering the JVM instead of copies of the application, Terracotta simplifies and streamlines many aspects of configuring for application performance. It allows more servers to be added to the machine cluster as traffic builds.
"They offload functionality that the database typically gets bogged down with," said Brad Shimmin, an analyst with Current Analysis. And Terracotta's approach "doesn't require the developer to do anything." Terracotta, on its 2.6 release since Monday, gets added to the application infrastructure, monitors the functioning Java Virtual Machines, and manages their data and object traffic for a server cluster.
"It's not a case of the developer saying, 'Oh, my God, I've got to embed another Java library in my application'," said Shimmin.
Terracotta is doing the opposite of the server consolidation that is going on elsewhere in the data center, said Amit Pandey, Terracotta CEO, in an interview. Instead of subdividing a physical machine by adding applications, Terracotta 2.6 is multiplying JVMs to make one server handle more traffic for one application. As in other forms of clustering, the Terracotta approach allows a JVM to fail without losing its transactions or data, shifting them over to another operative JVM.
Terracotta keeps track of objects and data in server memory and shares that information across a server cluster. Its function in some ways is duplicated by an in-memory database, such as Oracle's TimesTen, that can sit in front of a disk-based system and speed performance. Given its data tracking abilities, it also resembles Oracle Tangosol, which links up server memories into a grid and manages data for the group.
But Terracotta's expertise in clustering Java Virtual Machines springs from the experience of Ari Zilka, a co-founder and chief technology officer of the San Francisco firm, and veteran of scaling up Wal-Mart's Web site operations a few years ago. Unlike Oracle, Terracotta 2.6 is also open source code. It has been deployed by Comcast, Pearson VUE, PartyPoker, and other large-scale Java application users.
Pandey says he knows of 150 deployments, most of which have occurred over the past year as the product became open source code. Terracotta is available in commercially supported versions for $2,000 to $10,000 per node, with deals averaging $7,000 per node, Pandey said.
Microsoft Releases Dynamics CRM Online
Microsoft's been dabbling with on-demand software services for awhile, but the general availability of Microsoft Dynamics CRM Online, beginning Tuesday, marks its most significant effort yet to provide the market with an alternative to Salesforce.com.
The software service for managing a business's customer contacts, sales information, and marketing efforts is available as a subscription and hosted from Microsoft's data centers using a multi-tenant architecture. General availability follows months of testing by 500 Microsoft customers.
Microsoft is trying to beat Salesforce.com on price and storage options. The base version, called Professional, costs $44 per month per user following a one-year introductory rate of $39 per month. That includes 5 GB of data storage per organization and the ability to customize workflows. The Professional Plus version costs $59 per user per month with 20 Gbytes of storage, with more customization features and the ability to synchronize data contained in other systems with the service. Users access the service using Microsoft Outlook or a Web browser as an interface.
By comparison, Salesforce.com's Professional edition starts at $65 per month per user and comes with 1 Gbyte storage per organization; additional storage costs extra. While many small or midsize businesses won't even need the minimum of 5 Gbytes of storage offered by Microsoft, "this goes back to our strategy to create a new price-to-value equation," said Bill Patterson, group product manager of Dynamics CRM.
Salesforce.com has long argued that Dynamics CRM -- previously only available as an online subscription through resellers (called Dynamics CRM Live) or as a licensed, on-premise software application -- lacks features and functionality offered in its online software. Research firm Gartner ranked Salesfore.com as a leader in its magic quadrant for Salesforce Automation last year, a status it shared only with Oracle Siebel CRM, while Dynamics CRM was ranked as a challenger.
While Dynamics CRM may appeal to Microsoft-centric companies, it lacks some "best-of-breed" functionality and "proof points" that it can integrate well with companies' SAP and Oracle ERP systems, according to Gartner. That report was based on the analysis of Microsoft's software prior to an upgrade, Dynamics CRM 4.0, that became generally available in January.
Microsoft's Patterson argues the company trumps Salesforce.com on some features. Its Professional version, for example, offers 100 custom entities -- or objects for defining custom data forms, views, and attributes -- and 100 custom workflows, while Professional Plus provides 200 custom entities and 200 workflows. "By comparison, Salesforce.com gives you 50 custom entities and no custom workflows" for the Professional edition, he said. Salesforce.com does offer custom workflows in its Enterprise and Unlimited Editions, starting at $125 per user per month.
The software service for managing a business's customer contacts, sales information, and marketing efforts is available as a subscription and hosted from Microsoft's data centers using a multi-tenant architecture. General availability follows months of testing by 500 Microsoft customers.
Microsoft is trying to beat Salesforce.com on price and storage options. The base version, called Professional, costs $44 per month per user following a one-year introductory rate of $39 per month. That includes 5 GB of data storage per organization and the ability to customize workflows. The Professional Plus version costs $59 per user per month with 20 Gbytes of storage, with more customization features and the ability to synchronize data contained in other systems with the service. Users access the service using Microsoft Outlook or a Web browser as an interface.
By comparison, Salesforce.com's Professional edition starts at $65 per month per user and comes with 1 Gbyte storage per organization; additional storage costs extra. While many small or midsize businesses won't even need the minimum of 5 Gbytes of storage offered by Microsoft, "this goes back to our strategy to create a new price-to-value equation," said Bill Patterson, group product manager of Dynamics CRM.
Salesforce.com has long argued that Dynamics CRM -- previously only available as an online subscription through resellers (called Dynamics CRM Live) or as a licensed, on-premise software application -- lacks features and functionality offered in its online software. Research firm Gartner ranked Salesfore.com as a leader in its magic quadrant for Salesforce Automation last year, a status it shared only with Oracle Siebel CRM, while Dynamics CRM was ranked as a challenger.
While Dynamics CRM may appeal to Microsoft-centric companies, it lacks some "best-of-breed" functionality and "proof points" that it can integrate well with companies' SAP and Oracle ERP systems, according to Gartner. That report was based on the analysis of Microsoft's software prior to an upgrade, Dynamics CRM 4.0, that became generally available in January.
Microsoft's Patterson argues the company trumps Salesforce.com on some features. Its Professional version, for example, offers 100 custom entities -- or objects for defining custom data forms, views, and attributes -- and 100 custom workflows, while Professional Plus provides 200 custom entities and 200 workflows. "By comparison, Salesforce.com gives you 50 custom entities and no custom workflows" for the Professional edition, he said. Salesforce.com does offer custom workflows in its Enterprise and Unlimited Editions, starting at $125 per user per month.
Green Enterprise Class Data Center from Hitachi Data Systems
Breaking new ground in data center design and development, Hitachi Data Systems, a wholly owned subsidiary of Hitachi, has announced a eco-friendly and power-efficient data center design which achieves a 1.6 PUE (Power Usage Effectiveness) rating by The Green Grid. This is the lowest power usage index of any data center in its class. Located in Yokohama, Japan, the new state-of-the-art data center brings together the collective product innovations of the Hitachi Information & Technology Systems Group (ITSG) companies including industry-leading storage systems, server and networking equipment. This sophisticated architecture is designed to offer the highest levels of energy efficiency and reduce carbon emissions by 20 percent, while also lowering IT management costs.
Optimized to deliver even greater power efficiencies, the facility also features cutting-edge products such as Thermal hydraulic cooling devices, uninterruptible power supply systems and highly advanced power supply converters which will work in concert to contribute significant power, cooling and space benefits throughout. Leading-edge patented Hitachi Finger Vein authentication and other RFID technologies have also been integrated into the facility for state-of-the-art security.
"Within the IT industry, the Wikibon community believes that Hitachi has the most comprehensive and fully implemented corporate green plan in place," says David Vellante, president and CEO of IT Centrix and co-founder of the Wikibon Project. "Their progress on its emission neutral strategy is impressive and genuine. Initiatives such as the collaboration of various Hitachi groups for a new datacenter design in Yokohama underscore the firm’s commitment and are great drivers for change. Within storage the USP V controller re-designs and the implementation of virtualization, thin provisioning and support for external devices that spin down, have helped improve utilization and reduce power consumption by 63% over previous generations, a substantial milestone that sets an example of leadership for the industry."
Environmental issues continue to escalate IT budget discussions, impact data center designs, and shape corporate social responsibility. The new data center marks another chief milestone in support of Hitachi’s Harmonious Green Plan and Project CoolCenter50 corporate initiatives which are aimed at reducing 330,000 tons of carbon emissions and cutting down power consumption by as much as 50 percent by 2012. It is the company’s charter to apply these targets to the product development process across the entire IT solutions portfolio—chief among which include Hitachi Data Systems’ services-oriented storage solutions.
“In today’s climate, we have found that the cost of energy and power has quickly emerged as a growing concern, driving companies to take a deeper look at the efficiency of their data centers,” said Hu Yoshida, vice president and CTO, Hitachi Data Systems. “IT organizations need to take a holistic approach and carefully examine how every facet of their data center can play a role in improving their environmental impact—and lowering escalating power consumption levels. By leveraging Hitachi green technology, this breakthrough data center architecture provides a best practices approach for driving better efficiencies and utilization in customer environments now and in the future.”
Optimized to deliver even greater power efficiencies, the facility also features cutting-edge products such as Thermal hydraulic cooling devices, uninterruptible power supply systems and highly advanced power supply converters which will work in concert to contribute significant power, cooling and space benefits throughout. Leading-edge patented Hitachi Finger Vein authentication and other RFID technologies have also been integrated into the facility for state-of-the-art security.
"Within the IT industry, the Wikibon community believes that Hitachi has the most comprehensive and fully implemented corporate green plan in place," says David Vellante, president and CEO of IT Centrix and co-founder of the Wikibon Project. "Their progress on its emission neutral strategy is impressive and genuine. Initiatives such as the collaboration of various Hitachi groups for a new datacenter design in Yokohama underscore the firm’s commitment and are great drivers for change. Within storage the USP V controller re-designs and the implementation of virtualization, thin provisioning and support for external devices that spin down, have helped improve utilization and reduce power consumption by 63% over previous generations, a substantial milestone that sets an example of leadership for the industry."
Environmental issues continue to escalate IT budget discussions, impact data center designs, and shape corporate social responsibility. The new data center marks another chief milestone in support of Hitachi’s Harmonious Green Plan and Project CoolCenter50 corporate initiatives which are aimed at reducing 330,000 tons of carbon emissions and cutting down power consumption by as much as 50 percent by 2012. It is the company’s charter to apply these targets to the product development process across the entire IT solutions portfolio—chief among which include Hitachi Data Systems’ services-oriented storage solutions.
“In today’s climate, we have found that the cost of energy and power has quickly emerged as a growing concern, driving companies to take a deeper look at the efficiency of their data centers,” said Hu Yoshida, vice president and CTO, Hitachi Data Systems. “IT organizations need to take a holistic approach and carefully examine how every facet of their data center can play a role in improving their environmental impact—and lowering escalating power consumption levels. By leveraging Hitachi green technology, this breakthrough data center architecture provides a best practices approach for driving better efficiencies and utilization in customer environments now and in the future.”
World's Hottest Tech Markets
In an ominous sign for Microsoft's growth prospects in emerging regions, countries that represent the world's fastest growing tech markets voted against accepting the company's latest Microsoft Office document format as an international standard.
Brazil, India, and China, which together count for more than a third of the world's population, all voted against Office Open XML in voting last week before the International Organization for Standardization (ISO). Russia was the only member of the so-called BRIC nations to vote in favor of ISO ratification for OOXML.
"These are areas where open source has more strength and more advocates," said Gartner analyst Michael Silver in an interview.
Seventy-five percent of ISO member nations voted in favor of making OOXML an ISO standard -- meaning that the format has won the standards body's imprimatur, barring a successful appeal by opponents.
But the list of large country's that voted against OOXML also includes Canada, Iran, South Africa and Venezuela. The United States voted in favor of the format, which is used in Microsoft's Office 2007 programs.
Developing nations like China, India, and Brazil are expected to account for the bulk of tech spending growth in the coming years. The fact that those countries rejected OOXML as a standard should be troubling for Microsoft, Silver said.
"Microsoft has long been trying to figure out the best ways to get into these countries, but those are areas where users are looking for more open source and free products," said Silver.
Silver said the market share for the Linux operating system and applications that use the Open Document Format is small but growing fast in emerging markets, where consumers and businesses lack large technology budgets. "They're much more price sensitive and don't have a pre-existing installed base that locks them into commercial products," said Silver.
Gartner estimates that the market share for desktop Linux is about 2% in the Asia-Pacific region, 4.5% in Eastern Europe, and 4% in the Middle East and Africa. By contrast, Linux holds 1.2% of the desktop market in the U.S.
Microsoft's apparent victory before the ISO could be subject to appeal.
ISO approval of OOXML comes amid widespread allegations that Microsoft improperly tried to influence voting and the EU is investigating the process.
The chairman of a Norwegian technology committee tasked with studying OOXML earlier this week filed a protest against his country's decision to approve the format. Questions have also been raised about the voting process in Germany, Croatia, and several other countries.
Microsoft last year conceded that an employee in Sweden offered to compensate local tech execs for joining the country's standards committee and voting in favor of OOXML.
The stakes are high. OOXML is the default file format for Microsoft Office 2007 applications, including Word, PowerPoint, and Excel. The ISO recognition of OOXML could open the door to lucrative government, non-profit and educational markets for Office 2007.
Critics have argued that Microsoft has failed to publish sufficient documentation about OOXML for it to be considered a truly open standard.
ISO member nations last year rejected Microsoft's initial request for OOXML approval. National bodies last week wrapped up voting on the company's follow-up request.
Brazil, India, and China, which together count for more than a third of the world's population, all voted against Office Open XML in voting last week before the International Organization for Standardization (ISO). Russia was the only member of the so-called BRIC nations to vote in favor of ISO ratification for OOXML.
"These are areas where open source has more strength and more advocates," said Gartner analyst Michael Silver in an interview.
Seventy-five percent of ISO member nations voted in favor of making OOXML an ISO standard -- meaning that the format has won the standards body's imprimatur, barring a successful appeal by opponents.
But the list of large country's that voted against OOXML also includes Canada, Iran, South Africa and Venezuela. The United States voted in favor of the format, which is used in Microsoft's Office 2007 programs.
Developing nations like China, India, and Brazil are expected to account for the bulk of tech spending growth in the coming years. The fact that those countries rejected OOXML as a standard should be troubling for Microsoft, Silver said.
"Microsoft has long been trying to figure out the best ways to get into these countries, but those are areas where users are looking for more open source and free products," said Silver.
Silver said the market share for the Linux operating system and applications that use the Open Document Format is small but growing fast in emerging markets, where consumers and businesses lack large technology budgets. "They're much more price sensitive and don't have a pre-existing installed base that locks them into commercial products," said Silver.
Gartner estimates that the market share for desktop Linux is about 2% in the Asia-Pacific region, 4.5% in Eastern Europe, and 4% in the Middle East and Africa. By contrast, Linux holds 1.2% of the desktop market in the U.S.
Microsoft's apparent victory before the ISO could be subject to appeal.
ISO approval of OOXML comes amid widespread allegations that Microsoft improperly tried to influence voting and the EU is investigating the process.
The chairman of a Norwegian technology committee tasked with studying OOXML earlier this week filed a protest against his country's decision to approve the format. Questions have also been raised about the voting process in Germany, Croatia, and several other countries.
Microsoft last year conceded that an employee in Sweden offered to compensate local tech execs for joining the country's standards committee and voting in favor of OOXML.
The stakes are high. OOXML is the default file format for Microsoft Office 2007 applications, including Word, PowerPoint, and Excel. The ISO recognition of OOXML could open the door to lucrative government, non-profit and educational markets for Office 2007.
Critics have argued that Microsoft has failed to publish sufficient documentation about OOXML for it to be considered a truly open standard.
ISO member nations last year rejected Microsoft's initial request for OOXML approval. National bodies last week wrapped up voting on the company's follow-up request.
Gartner Cites Six Best Practices In Virtualizing Servers
The biggest change underway in the data center is virtualization, said two Gartner IT analysts this week.
Gartner analysts have had "a thousand conversations" with clients on the subject, and as a result, they have come up with a list of six best practices for server virtualization.
Consultants may recommend large-scale server virtualization. Gartner said start small and achieve the concrete server consolidation that you're seeking in a first phase." The second phase is more strategically important, more complex to implement and provides far more value for the customer," say Thomas Bittman and John Enck in comments accompanying the list. The second phase concentrates on the flexible allocation of resources to respond to business demand, such as starting up more virtual machines or assigning more resources to a virtual machine with a priority task. "In this phase, the focus shifts to delivering new services or improving the quality and speed of service," they wrote. In other words, start small, but think big, they say.
Their second point is "Require a rapid return on investment." The IT organization "needs to build a business case with a rapid return on investment" because the virtualization market is evolving rapidly, Bittman and Enck write. They recommend shooting for an ROI within six months or less, although they don't indicate whether they think prices are likely to go up or down in the rapidly evolving market. Nevertheless, their benchmark is: deploying a minimum of 50 virtual machines in a year will make good on the investment in virtualization software.
"Virtualize the right applications," they urged. Existing applications that utilize most of the hardware resources on which they sit "are not going to generate savings" by migrating them into virtual machines, they warned. Applications with high input/output traffic may become inefficient on virtual machines, because in most cases multiple virtual machines are sharing the limited I/O capacity of one piece of hardware. "Older, smaller packaged applications" make the best target for initial virtualization. The majority of such applications being moved into virtual machines are none the less deployed in production "in mission critical roles."
Define a storage strategy to go with your server virtualization plans, they noted. The images of the virtual machines on which your organization depends need to be always accessible. If they are stored on disks directly attached to the server, then the failure of that server or storage system will make the images inaccessible. Putting them on a central storage system gives the enterprise "the flexibility to access virtual images from any server connected to the storage system," they pointed out.
"Understand software licensing issues," they advised. Virtualization has appeared on the scene before independent software vendors have thought through how they want to charge for their products, once they're used in virtual machines. Keeping prices low would encourage usage in the midst of a rush top virtualize. But it's tempting also to just collect the added license fees as what used to be one copy of the product becomes three or four or more in virtual machines. "Gartner predicts that software pricing and licensing will remain problematic for the near future," said Bittman and Enck. That means don't expect a price break just because you want to increase usage of an ISV's product. But virtualization remains the "the most important trend through 2012." As competition sharpens, ISVs may adjust their pricing to emerge as the big benefactors of the trend. "Until new pricing models are found, users should seek to understand ISV's pricing and licensing policies in as much detail as possible."
"Combine virtual machines effectively." Server administrators will try to balance an application that has high I/O during the workday with one that utilizes I/O channels for backup at night. But Bittman and Enck recommend trying to construct a dynamic way to balance workloads rather than devising "a perfect static consolidation mapping." The workloads are inevitably going to go up and down and may not be a perfect match at all times, if virtual machines achieve the 75-80% server utilization rate that many data center managers now consider optimum. "Being able to deal with these changes dynamically is a key goal, particularly in the early stages of virtualization," they wrote. That means buying into the management tools that VMware, Virtual Iron, XenSource and other third parties, such as Akorri and Veeam, now offer.
The market for virtualization software will mature over the next five years, but "most enterprises can't afford to wait--server sprawl, data center space and power problems are here now," wrote Enck. Following these guidelines, many organizations will sidestep these problems and proceed to virtualize their servers, he predicted.
"Many of these problems can be avoided if enterprises make the proper assessments before they virtualize their machines," he wrote.
Gartner analysts have had "a thousand conversations" with clients on the subject, and as a result, they have come up with a list of six best practices for server virtualization.
Consultants may recommend large-scale server virtualization. Gartner said start small and achieve the concrete server consolidation that you're seeking in a first phase." The second phase is more strategically important, more complex to implement and provides far more value for the customer," say Thomas Bittman and John Enck in comments accompanying the list. The second phase concentrates on the flexible allocation of resources to respond to business demand, such as starting up more virtual machines or assigning more resources to a virtual machine with a priority task. "In this phase, the focus shifts to delivering new services or improving the quality and speed of service," they wrote. In other words, start small, but think big, they say.
Their second point is "Require a rapid return on investment." The IT organization "needs to build a business case with a rapid return on investment" because the virtualization market is evolving rapidly, Bittman and Enck write. They recommend shooting for an ROI within six months or less, although they don't indicate whether they think prices are likely to go up or down in the rapidly evolving market. Nevertheless, their benchmark is: deploying a minimum of 50 virtual machines in a year will make good on the investment in virtualization software.
"Virtualize the right applications," they urged. Existing applications that utilize most of the hardware resources on which they sit "are not going to generate savings" by migrating them into virtual machines, they warned. Applications with high input/output traffic may become inefficient on virtual machines, because in most cases multiple virtual machines are sharing the limited I/O capacity of one piece of hardware. "Older, smaller packaged applications" make the best target for initial virtualization. The majority of such applications being moved into virtual machines are none the less deployed in production "in mission critical roles."
Define a storage strategy to go with your server virtualization plans, they noted. The images of the virtual machines on which your organization depends need to be always accessible. If they are stored on disks directly attached to the server, then the failure of that server or storage system will make the images inaccessible. Putting them on a central storage system gives the enterprise "the flexibility to access virtual images from any server connected to the storage system," they pointed out.
"Understand software licensing issues," they advised. Virtualization has appeared on the scene before independent software vendors have thought through how they want to charge for their products, once they're used in virtual machines. Keeping prices low would encourage usage in the midst of a rush top virtualize. But it's tempting also to just collect the added license fees as what used to be one copy of the product becomes three or four or more in virtual machines. "Gartner predicts that software pricing and licensing will remain problematic for the near future," said Bittman and Enck. That means don't expect a price break just because you want to increase usage of an ISV's product. But virtualization remains the "the most important trend through 2012." As competition sharpens, ISVs may adjust their pricing to emerge as the big benefactors of the trend. "Until new pricing models are found, users should seek to understand ISV's pricing and licensing policies in as much detail as possible."
"Combine virtual machines effectively." Server administrators will try to balance an application that has high I/O during the workday with one that utilizes I/O channels for backup at night. But Bittman and Enck recommend trying to construct a dynamic way to balance workloads rather than devising "a perfect static consolidation mapping." The workloads are inevitably going to go up and down and may not be a perfect match at all times, if virtual machines achieve the 75-80% server utilization rate that many data center managers now consider optimum. "Being able to deal with these changes dynamically is a key goal, particularly in the early stages of virtualization," they wrote. That means buying into the management tools that VMware, Virtual Iron, XenSource and other third parties, such as Akorri and Veeam, now offer.
The market for virtualization software will mature over the next five years, but "most enterprises can't afford to wait--server sprawl, data center space and power problems are here now," wrote Enck. Following these guidelines, many organizations will sidestep these problems and proceed to virtualize their servers, he predicted.
"Many of these problems can be avoided if enterprises make the proper assessments before they virtualize their machines," he wrote.
SaaS, Web Services
Business and IT managers have ranked software-as-a-service and Web services as the most important trends in the software industry for the third year in a row, according to an annual survey by McKinsey & Co. and Sand Hill Group, to be presented next week at the Interop/Software 2008 conference in Las Vegas.
In a survey of 857 managers, 23% ranked SaaS as the most important item for their businesses in 2008, up slightly from 21% last year, but down from 30% in 2006. One in four respondents ranked Web services/SOA as the most important, up from 18% in 2007 and 24% in 2006. Trailing those issues were open source software, offshore outsourcing, and software industry consolidation.
Yet SaaS continues to appeal to small businesses. Among companies with between 1,000 and 25,000 employees, an average of 11% of software budgets were spent on SaaS, with 70% or more of budgets going to traditional software licenses and maintenance. In contrast, respondents with fewer than 100 employees spent 26% of their budgets on SaaS, while those with 100-1,000 employees spent 17%. Among companies with under $1 billion in annual revenues, 46% had purchased at least one SaaS application.
The McKinsey's research also revealed some maturation of SaaS among small businesses. Thirty-six percent of small and midsize businesses all respondents had adopted are using multiple SaaS applications, and eight out of ten of those had bought multiple SaaS applications. Only 12% of respondents said they had adopted their first SaaS application in 2007 the last year, compared with one-third of respondents who adopted their first SaaS app in 2007 2006 who had adopted their first SaaS app the previous year.
"Peak adoption happened in 2006, and now it's a question of deeper penetration of SaaS," said Junaid Mohiuddin, a software consultant at McKinsey & Co.
Yet not all of those SaaS purchases were software served up through a service.
McKinsey and Sand Hill took a broad view of SaaS in their research, including such things as online storage and security services. Respondents ranked online storage, in fact, as their most commonly used SaaS application, followed by online backup, security services, system and network management, customer-relationship management, and collaboration software, respectively.
The No. 1 ranked criteria for vendor selection of SaaS was deployment speed and ease of integration, followed by the vendor's track record in SaaS, and costs.
In a survey of 857 managers, 23% ranked SaaS as the most important item for their businesses in 2008, up slightly from 21% last year, but down from 30% in 2006. One in four respondents ranked Web services/SOA as the most important, up from 18% in 2007 and 24% in 2006. Trailing those issues were open source software, offshore outsourcing, and software industry consolidation.
Yet SaaS continues to appeal to small businesses. Among companies with between 1,000 and 25,000 employees, an average of 11% of software budgets were spent on SaaS, with 70% or more of budgets going to traditional software licenses and maintenance. In contrast, respondents with fewer than 100 employees spent 26% of their budgets on SaaS, while those with 100-1,000 employees spent 17%. Among companies with under $1 billion in annual revenues, 46% had purchased at least one SaaS application.
The McKinsey's research also revealed some maturation of SaaS among small businesses. Thirty-six percent of small and midsize businesses all respondents had adopted are using multiple SaaS applications, and eight out of ten of those had bought multiple SaaS applications. Only 12% of respondents said they had adopted their first SaaS application in 2007 the last year, compared with one-third of respondents who adopted their first SaaS app in 2007 2006 who had adopted their first SaaS app the previous year.
"Peak adoption happened in 2006, and now it's a question of deeper penetration of SaaS," said Junaid Mohiuddin, a software consultant at McKinsey & Co.
Yet not all of those SaaS purchases were software served up through a service.
McKinsey and Sand Hill took a broad view of SaaS in their research, including such things as online storage and security services. Respondents ranked online storage, in fact, as their most commonly used SaaS application, followed by online backup, security services, system and network management, customer-relationship management, and collaboration software, respectively.
The No. 1 ranked criteria for vendor selection of SaaS was deployment speed and ease of integration, followed by the vendor's track record in SaaS, and costs.
AMD's Three More Phenom Triple-Core Processors
Advanced Micro Devices on Wednesday introduced three triple-core Phenom desktop processors for less than $200.
Combined with the AMD 780 series chipsets, the new chips are aimed at high-end PCs for entertainment and games. For additional power, an ATI Radeon graphics process from AMD could be added. The card's "hybrid graphics technology" makes it possible to run the graphics processor in conjunction with processor on the chipset to give PC performance an added jolt, AMD said.
The three new models are the Phenom X3 8750, 8650 and 8450, which have clock speeds of 2.4 GHz, 2.3 GHz and 2.1 GHz, respectively, per core. Pricing is $195 for the 8750, $165 for the $8650 and $145 for the 8450.
AMD's strategy behind its triple-core product line is to offer a processor that's faster than dual-core products, while less expensive than the quad-core processors offered by rival Intel. While triple-core chips are not as powerful as quad-core processors, the former deliver the power needed for mainstream entertainment and gaming PCs at a lower price.
AMD introduced the Phenom X3 8000 series last month with the 8600 and 8400 models. Pricing was about $150 for the 2.1 GHz 8400 and about $175 for the 2.3 GHz 8600.
In pairing the 8000 series with its 780 chipsets, AMD is offering a platform that it claims provides smooth HD viewing of movies in Blu-ray format. Supported video-related standards include H.264 for video compression, MPEG-2 for the generic coding of moving pictures and associated audio information, and VC-1, a video codec standard initially developed by Microsoft.
Combined with the AMD 780 series chipsets, the new chips are aimed at high-end PCs for entertainment and games. For additional power, an ATI Radeon graphics process from AMD could be added. The card's "hybrid graphics technology" makes it possible to run the graphics processor in conjunction with processor on the chipset to give PC performance an added jolt, AMD said.
The three new models are the Phenom X3 8750, 8650 and 8450, which have clock speeds of 2.4 GHz, 2.3 GHz and 2.1 GHz, respectively, per core. Pricing is $195 for the 8750, $165 for the $8650 and $145 for the 8450.
AMD's strategy behind its triple-core product line is to offer a processor that's faster than dual-core products, while less expensive than the quad-core processors offered by rival Intel. While triple-core chips are not as powerful as quad-core processors, the former deliver the power needed for mainstream entertainment and gaming PCs at a lower price.
AMD introduced the Phenom X3 8000 series last month with the 8600 and 8400 models. Pricing was about $150 for the 2.1 GHz 8400 and about $175 for the 2.3 GHz 8600.
In pairing the 8000 series with its 780 chipsets, AMD is offering a platform that it claims provides smooth HD viewing of movies in Blu-ray format. Supported video-related standards include H.264 for video compression, MPEG-2 for the generic coding of moving pictures and associated audio information, and VC-1, a video codec standard initially developed by Microsoft.
Latest Update 24-4-08
Security vendor McAfee has plans to invest $30 million in its Indian operations to spruce up its R&D activity and strengthen its channels.
The company’s India Development Center was opened in Bangalore in 2002 and has been responsible for creating its online anti-virus solution which boasts of over 5 million customers globally. The bulk of the proposed funds will be directed towards increasing its R&D activities.
In January 2008, the company had launched a Preferred Partner Program, where in addition to the profit margin on the product sold, the company is offering a rebate as well.
“Since inception, we have invested over $80 million in our India operations and plan to add $30 million to that this year. With almost 25 percent of our global employees located in India, we are committed to our operations here. We will increase our head count for our R&D activities and will look at adding new functionality to the mid-market product range,” said Darrell Rodenbaugh, Senior VP, Global Mid-Market Business, McAfee.
Through its recently launched program, the company hopes to improve its channel relationships by offering priority support to those partners who have signed up in sales and tech support. Currently, they have signed on 30 partners, and plan to reach 100 by the end of the year. As part of this program, these partners will have access to a marketing fund which will help them tap opportunities.
Rodenbaugh added,“We are growing at over 30 percent year-on-year and we need dedicated support from our channel partners to sustain it. With this program, we will expect our partners to have a dedicated approach for selling our product line in the country.”
The company’s India Development Center was opened in Bangalore in 2002 and has been responsible for creating its online anti-virus solution which boasts of over 5 million customers globally. The bulk of the proposed funds will be directed towards increasing its R&D activities.
In January 2008, the company had launched a Preferred Partner Program, where in addition to the profit margin on the product sold, the company is offering a rebate as well.
“Since inception, we have invested over $80 million in our India operations and plan to add $30 million to that this year. With almost 25 percent of our global employees located in India, we are committed to our operations here. We will increase our head count for our R&D activities and will look at adding new functionality to the mid-market product range,” said Darrell Rodenbaugh, Senior VP, Global Mid-Market Business, McAfee.
Through its recently launched program, the company hopes to improve its channel relationships by offering priority support to those partners who have signed up in sales and tech support. Currently, they have signed on 30 partners, and plan to reach 100 by the end of the year. As part of this program, these partners will have access to a marketing fund which will help them tap opportunities.
Rodenbaugh added,“We are growing at over 30 percent year-on-year and we need dedicated support from our channel partners to sustain it. With this program, we will expect our partners to have a dedicated approach for selling our product line in the country.”
This virtual museum of Computing
This virtual museum includes an eclectic collection of World Wide Web (WWW) hyperlinks connected with the history of computing and on-line computer-based exhibits available both locally and around the world. It was founded on 1 June 1995, so is an example of an "old" virtual museum itself. [VISITOR NUMBER]
This museum opened on 1st June 1995. The museum receives about 200 visitors each day. Please contact Jonathan Bowen if you know of relevant on-line information not included here.
Mirror sites are available in Sweden and USA courtesy of ICOM, and also elsewhere, including the UK, if you experience poor access speed. Automatic redirection to a mirror site is available.
This museum opened on 1st June 1995. The museum receives about 200 visitors each day. Please contact Jonathan Bowen if you know of relevant on-line information not included here.
Mirror sites are available in Sweden and USA courtesy of ICOM, and also elsewhere, including the UK, if you experience poor access speed. Automatic redirection to a mirror site is available.
Wednesday, April 23, 2008
Human-centered computing
Human-centered computing (HCC) is an emerging, interdisciplinary academic field broadly concerned with computing and computational artifacts as they relate to the human condition. Researchers and practitioners who affiliate themselves with human-centered computing usually come from one or more of the following disciplines: computer science, sociology, psychology, cognitive science, anthropology, communication studies, science and technology studies, and industrial design.
Research in human-centered computing has multiple goals. Some researchers focus on understanding humans, both as individuals and in social groups, by focusing on the ways that human beings adopt, adapt, and organize their lives around computational technologies. Others focus on developing new design strategies for computational artifacts. Human-centered design of computational tools attempts to address problems that traditional design approaches, such as heuristic evaluations and measurements of productivity and efficiency, do not generally address. Designing computational tools for spirituality, for fun, and for leisure are some examples of non-traditional design problems that are of interest to HCC researchers and engineers. HCC researchers also bring a diverse array of conceptual and research tools to traditional computing areas such as computer-supported cooperative work, computer-supported collaborative learning, and ubiquitous computing.
Human-centered computing is closely related to other interdisciplinary fields such as human-computer interaction and information science, and exactly where the boundaries between these fields lie is not clear. Broadly speaking, however, human-centered computing usually concerns itself with systems and practices of technology use. Human-computer interaction is more focused on ergonomics and the usability of computing artifacts, while information science is focused on practices surrounding the collection, manipulation, and use of information.
Research in human-centered computing has multiple goals. Some researchers focus on understanding humans, both as individuals and in social groups, by focusing on the ways that human beings adopt, adapt, and organize their lives around computational technologies. Others focus on developing new design strategies for computational artifacts. Human-centered design of computational tools attempts to address problems that traditional design approaches, such as heuristic evaluations and measurements of productivity and efficiency, do not generally address. Designing computational tools for spirituality, for fun, and for leisure are some examples of non-traditional design problems that are of interest to HCC researchers and engineers. HCC researchers also bring a diverse array of conceptual and research tools to traditional computing areas such as computer-supported cooperative work, computer-supported collaborative learning, and ubiquitous computing.
Human-centered computing is closely related to other interdisciplinary fields such as human-computer interaction and information science, and exactly where the boundaries between these fields lie is not clear. Broadly speaking, however, human-centered computing usually concerns itself with systems and practices of technology use. Human-computer interaction is more focused on ergonomics and the usability of computing artifacts, while information science is focused on practices surrounding the collection, manipulation, and use of information.
Human-based computation as a form of social organization
Viewed as a form of social organization, human-based computation often surprisingly turns out to be more robust and productive than traditional organizations (Kosorukoff and Goldberg, 2002). The latter depend on obligations to maintain their more or less fixed structure, be functional and stable. Each of them is similar to a carefully designed mechanism with humans as its parts. However, this limits the freedom of their human employees and subjects them to various kinds of stresses. Most people, unlike mechanical parts, find it difficult to adapt to some fixed roles that best fit the organization. Evolutionary human-computation projects offer a natural solution to this problem. They adapt organizational structure to human spontaneity, accommodate human mistakes and creativity, and utilize both in a constructive way. This leaves their participants free from obligations without endangering the functionality of the whole, making people happier. There are still some challenging research problems that need to be solved before we can realize the full potential of this idea.
The algorithmic outsourcing techniques used in human-based computation are much more scalable than the manual or automated techniques used to manage outsourcing traditionally. It is this scalability that allows to easily distribute the effort among thousands of participants. It was suggested recently that this mass outsourcing is sufficiently different from traditional small-scale outsourcing to merit a new name crowdsourcing (Howe, 2006).
The algorithmic outsourcing techniques used in human-based computation are much more scalable than the manual or automated techniques used to manage outsourcing traditionally. It is this scalability that allows to easily distribute the effort among thousands of participants. It was suggested recently that this mass outsourcing is sufficiently different from traditional small-scale outsourcing to merit a new name crowdsourcing (Howe, 2006).
Classes of human-based computation
Human-based computation methods combine computers and humans in different roles. Kosorukoff (2000) proposed a way to describe division of labor in computation, that groups human-based methods into three classes. The following table uses the evolutionary computation model to describe four classes of computation, three of which rely on humans in some role. For each class, a representative example is shown. The classification is in terms of the roles (innovation or selection) performed in each case by humans and computational processes. This table is a slice of three-dimensional table. The third dimension defines if the organizational function is performed by humans or a computer. Here it is assumed to be performed by a computer.
Classes of human-based computation from this table can be referred by two-letter abbreviations: HC, CH, HH. Here the first letter identifies the type of agents performing innovation, the second letter specifies the type of selection agents. In some implementations (wiki is the most common example), human-based selection functionality might be limited, it can be shown with small h
| Division of labor in computation | |||
|---|---|---|---|
| Selection agent Innovation agent | Computer | Human | |
| Computer | Genetic Algorithm | Interactive genetic algorithm | |
| Human | Computerized Tests | Human-based genetic algorithm | |
Classes of human-based computation from this table can be referred by two-letter abbreviations: HC, CH, HH. Here the first letter identifies the type of agents performing innovation, the second letter specifies the type of selection agents. In some implementations (wiki is the most common example), human-based selection functionality might be limited, it can be shown with small h
Human-based computation
In computer science, human-based computation is a technique when a computational process performs its function via outsourcing certain steps to humans (Kosorukoff, 2001). This approach leverages differences in abilities and alternative costs between humans and computer agents to achieve symbiotic human-computer interaction.
In traditional computation, a human employs a computer to solve a problem: a human provides a formalized problem description to a computer, and receives a solution to interpret. In human-based computation, the roles are often reversed: the computer asks a person or a large number of people to solve a problem, then collects, interprets, and integrates their solutions.
In traditional computation, a human employs a computer to solve a problem: a human provides a formalized problem description to a computer, and receives a solution to interpret. In human-based computation, the roles are often reversed: the computer asks a person or a large number of people to solve a problem, then collects, interprets, and integrates their solutions.
Adiabatic quantum computation
Adiabatic Quantum Computation relies on the adiabatic theorem to do calculations.[1] First, a complex Hamiltonian is found whose ground state describes the solution to the problem of interest. Next, a system with a simple Hamiltonian is prepared and initialized to the ground state. Finally, the simple Hamiltonian is adiabatically evolved to the complex Hamiltonian. By the adiabatic theorem, the system remains in the ground state, so at the end the state of the system describes the solution to the problem.
Peptide computing
Peptide computing is a form of computing which uses peptides and molecular biology, instead of traditional silicon-based computer technologies. The basis of this computational model is the affinity of antibodies towards peptide sequences. Similar to DNA computing, the parallel interactions of peptide sequences and antibodies have been used by this model to solve a few NP-complete problems. Specifically, the hamiltonian path problem (HPP) and some versions of the set cover problem are a few NP-complete problems which have been solved using this computational model so far. This model of computation has also been shown to be computationally universal (or Turing complete).
This model of computation has some critical advantages over DNA computing. For instance, while DNA is made of four building blocks, peptides are made of twenty building blocks. The peptide-antibody interactions are also more flexible with respect to recognition and affinity than an interaction between a DNA strand and its reverse complement. However, unlike DNA computing, this model is yet to be practically realized. The main limitation is the availability of specific monoclonal antibodies required by the model.
This model of computation has some critical advantages over DNA computing. For instance, while DNA is made of four building blocks, peptides are made of twenty building blocks. The peptide-antibody interactions are also more flexible with respect to recognition and affinity than an interaction between a DNA strand and its reverse complement. However, unlike DNA computing, this model is yet to be practically realized. The main limitation is the availability of specific monoclonal antibodies required by the model.
Edge computing
Edge computing provides application processing load balancing capacity to corporate and other large-scale web servers. It is like an application cache, where the cache is in the Internet itself. Static web-sites being cached on mirror sites is not a new concept and Akamai has been fundamental to providing the mirroring and routing infrastructure to make that possible. Mirroring transactional and interactive systems are however a much more complex endeavor.
Overview
As the name implies, Edge computing pushes applications, data and computing power (services) away from centralized points to the logical extremes of a network. Edge computing replicates fragments of information across distributed networks of web servers, which may be vast and include many networks. As a topological paradigm, Edge computing is also referred to as mesh computing, peer-to-peer computing, autonomic (self-healing) computing, grid computing, and other names implying non-centralized, nodeless availability.
To ensure acceptable performance of widely-dispersed distributed services, large organizations typically implement Edge computing by deploying Web server farms with clustering. Previously available only to very large corporate and government organizations, technology advancement and cost reduction for large-scale implementations have made the technology available to small and medium-sized business.
The target end-user is any Internet client making use of commercial Internet application services.
Edge computing imposes certain limitations on the choices of technology platforms, applications or services, all of which need to be specifically developed or configured for edge computing.
Edge computing has many advantages:
1. Edge application services significantly decrease the data volume that must be moved, the consequent traffic, and the distance the data must go, thereby reducing transmission costs, shrinking latency, and improving quality of service (QoS).
2. Edge computing eliminates, or at least de-emphasizes, the core computing environment, limiting or removing a major bottleneck and a potential point of failure.
3. Security is also improved as encrypted data moves further in, toward the network core. As it approaches the enterprise, the data is checked as it passes through protected firewalls and other security points, where viruses, compromised data, and active hackers can be caught early on.
4. Finally, the ability to "virtualize" (i.e., logically group CPU capabilities on an as-needed, real-time basis) extends scalability. The Edge computing market is generally based on a "charge for network services" model, and it could be argued that typical customers for Edge services are organizations desiring linear scale of business application performance to the growth of, e.g., a subscriber base.
Overview
As the name implies, Edge computing pushes applications, data and computing power (services) away from centralized points to the logical extremes of a network. Edge computing replicates fragments of information across distributed networks of web servers, which may be vast and include many networks. As a topological paradigm, Edge computing is also referred to as mesh computing, peer-to-peer computing, autonomic (self-healing) computing, grid computing, and other names implying non-centralized, nodeless availability.
To ensure acceptable performance of widely-dispersed distributed services, large organizations typically implement Edge computing by deploying Web server farms with clustering. Previously available only to very large corporate and government organizations, technology advancement and cost reduction for large-scale implementations have made the technology available to small and medium-sized business.
The target end-user is any Internet client making use of commercial Internet application services.
Edge computing imposes certain limitations on the choices of technology platforms, applications or services, all of which need to be specifically developed or configured for edge computing.
Edge computing has many advantages:
1. Edge application services significantly decrease the data volume that must be moved, the consequent traffic, and the distance the data must go, thereby reducing transmission costs, shrinking latency, and improving quality of service (QoS).
2. Edge computing eliminates, or at least de-emphasizes, the core computing environment, limiting or removing a major bottleneck and a potential point of failure.
3. Security is also improved as encrypted data moves further in, toward the network core. As it approaches the enterprise, the data is checked as it passes through protected firewalls and other security points, where viruses, compromised data, and active hackers can be caught early on.
4. Finally, the ability to "virtualize" (i.e., logically group CPU capabilities on an as-needed, real-time basis) extends scalability. The Edge computing market is generally based on a "charge for network services" model, and it could be argued that typical customers for Edge services are organizations desiring linear scale of business application performance to the growth of, e.g., a subscriber base.
Examples of Social computing
Web 2.0
Main article: Web 2.0
A generation of internet applications was developed implementing aspects of social computing developed in the early 21st century.
Enterprise social software
Main article: Enterprise social software
Of particular interest in the realm of social computing is social software for enterprise. Sometimes referred to as "Enterprise 2.0",[2] a term derived from Web 2.0, this generally refers to the use of social computing in corporate intranets and in other medium- and large-scale business environments.
Electronic negotiation and electronic markets
Main article: Electronic negotiation
Electronic negotiation represents an important and desirable coordination mechanism for electronic markets. Negotiation between agents (software agents as well as humans) allows cooperatiove and competitive sharing of information to determine a proper price. Recent research and practice has also shown that electronic negotiation is beneficial for the coordination of complex interactions among organizations. Electronic negotiation has recently emerged as a very dynamic, interdisciplinary research area covering aspects from disciplines such as Economics, Information Systems, Computer Science, Communication Theory, Sociology and Psychology.
Collaborative filtering
Main article: Collaborative filtering
Collaborative filtering is the method of making automatic predictions (filtering) about the interests of a user by collecting taste information from many users (collaborating). Recommender systems often use it as a "social approach" in order to obtain music, movie, product, web site etc. recommendations.
Main article: Web 2.0
A generation of internet applications was developed implementing aspects of social computing developed in the early 21st century.
Enterprise social software
Main article: Enterprise social software
Of particular interest in the realm of social computing is social software for enterprise. Sometimes referred to as "Enterprise 2.0",[2] a term derived from Web 2.0, this generally refers to the use of social computing in corporate intranets and in other medium- and large-scale business environments.
Electronic negotiation and electronic markets
Main article: Electronic negotiation
Electronic negotiation represents an important and desirable coordination mechanism for electronic markets. Negotiation between agents (software agents as well as humans) allows cooperatiove and competitive sharing of information to determine a proper price. Recent research and practice has also shown that electronic negotiation is beneficial for the coordination of complex interactions among organizations. Electronic negotiation has recently emerged as a very dynamic, interdisciplinary research area covering aspects from disciplines such as Economics, Information Systems, Computer Science, Communication Theory, Sociology and Psychology.
Collaborative filtering
Main article: Collaborative filtering
Collaborative filtering is the method of making automatic predictions (filtering) about the interests of a user by collecting taste information from many users (collaborating). Recommender systems often use it as a "social approach" in order to obtain music, movie, product, web site etc. recommendations.
Rational Computing
Social computing begins with the observation that humans — and human behavior — are profoundly social. From birth humans orient to one another, and as they grow they develop abilities for interacting with one another ranging from expression and gesture to spoken and written language. As a consequence, people are remarkably sensitive to the behavior of those around them, and make countless decisions that are shaped by their social context. Whether it's wrapping up a talk when the audience starts fidgeting, choosing the crowded restaurant over the nearly deserted one, or crossing the street against the light because everyone else is doing so, social information provides a basis for inferences, planning, and coordinating activity.
The premise of social computing is that it is possible to design digital systems that support useful functionality by making socially produced information available to their users. This information may be provided directly, as when systems show the number of users who have rated a review as helpful or not. Or the information may be provided after being filtered and aggregated, as is done when systems recommend a product based on what else people with similar purchase history have purchased. Or the information may be provided indirectly, as is the case with Google’s page rank algorithms which orders search results based on the number of pages that (recursively) point to them. In all of these cases, information that is produced by a group of people is used to provide or enhance the functioning of a system. Social computing is concerned with systems of this sort and the mechanisms and principles that underlie them.
Social computing can be defined as follows:
Social Computing" refers to systems that support the gathering, representation, processing, use, and dissemination of information that is distributed across social collectivities such as teams, communities,organizations, and markets. Moreover, the information is not "anonymous" but is significant precisely because it is linked to people, who are in turn linked to other people.
The premise of social computing is that it is possible to design digital systems that support useful functionality by making socially produced information available to their users. This information may be provided directly, as when systems show the number of users who have rated a review as helpful or not. Or the information may be provided after being filtered and aggregated, as is done when systems recommend a product based on what else people with similar purchase history have purchased. Or the information may be provided indirectly, as is the case with Google’s page rank algorithms which orders search results based on the number of pages that (recursively) point to them. In all of these cases, information that is produced by a group of people is used to provide or enhance the functioning of a system. Social computing is concerned with systems of this sort and the mechanisms and principles that underlie them.
Social computing can be defined as follows:
Social Computing" refers to systems that support the gathering, representation, processing, use, and dissemination of information that is distributed across social collectivities such as teams, communities,organizations, and markets. Moreover, the information is not "anonymous" but is significant precisely because it is linked to people, who are in turn linked to other people.
Social computing
Social computing is a general term for an area of computer science that is concerned with the intersection of social behavior and computational systems. It is used in two ways.
In the weaker sense of the term, social computing has to do with supporting any sort of social behavior in or through computational systems. It is based on creating or recreating social conventions and social contexts through the use of software and technology. Thus, blogs, email, instant messaging, social network services, wikis, social bookmarking and other instances of what is often called social software illustrate ideas from social computing, but also auction software and other kinds of electronic market or electronic negotiation platforms where people interact socially.
In the stronger sense of the term, social computing has to do with supporting “computations” that are carried out by groups of people, an idea that has been popularized in James Surowiecki's book, The Wisdom of Crowds. Examples of social computing in this sense include collaborative filtering, online auctions, prediction markets, reputation systems, computational social choice, tagging, and verification games.
Social computing has become more widely known because of its relationship to a number of recent trends. These include the growing popularity of social software and Web 2.0, increased academic interest in social network analysis, the rise of open source as a viable method of production, and a growing conviction that all of this can have a profound impact on daily life. A February 13, 2006 paper by market research company Forrester Research suggested that:
“ Easy connections brought about by cheap devices, modular content, and shared computing resources are having a profound impact on our global economy and social structure. Individuals increasingly take cues from one another rather than from institutional sources like corporations, media outlets, religions, and political bodies. To thrive in an era of Social Computing, companies must abandon top-down management and communication tactics, weave communities into their products and services, use employees and partners as marketers, and become part of a living fabric of brand loyalists
In the weaker sense of the term, social computing has to do with supporting any sort of social behavior in or through computational systems. It is based on creating or recreating social conventions and social contexts through the use of software and technology. Thus, blogs, email, instant messaging, social network services, wikis, social bookmarking and other instances of what is often called social software illustrate ideas from social computing, but also auction software and other kinds of electronic market or electronic negotiation platforms where people interact socially.
In the stronger sense of the term, social computing has to do with supporting “computations” that are carried out by groups of people, an idea that has been popularized in James Surowiecki's book, The Wisdom of Crowds. Examples of social computing in this sense include collaborative filtering, online auctions, prediction markets, reputation systems, computational social choice, tagging, and verification games.
Social computing has become more widely known because of its relationship to a number of recent trends. These include the growing popularity of social software and Web 2.0, increased academic interest in social network analysis, the rise of open source as a viable method of production, and a growing conviction that all of this can have a profound impact on daily life. A February 13, 2006 paper by market research company Forrester Research suggested that:
“ Easy connections brought about by cheap devices, modular content, and shared computing resources are having a profound impact on our global economy and social structure. Individuals increasingly take cues from one another rather than from institutional sources like corporations, media outlets, religions, and political bodies. To thrive in an era of Social Computing, companies must abandon top-down management and communication tactics, weave communities into their products and services, use employees and partners as marketers, and become part of a living fabric of brand loyalists
Reconfigurable computing History
History and properties
The concept of reconfigurable computing has existed since the 1960s, when Gerald Estrin's landmark paper proposed the concept of a computer made of a standard processor and an array of “reconfigurable” hardware.[1][2] The main processor would control the behavior of the reconfigurable hardware. The latter would then be tailored to perform a specific task, such as image processing or pattern matching, as quickly as a dedicated piece of hardware. Once the task was done, the hardware could be adjusted to do some other task. This resulted in a hybrid computer structure combining the flexibility of software with the speed of hardware; unfortunately this idea was far ahead of its time in needed electronic technology.
In the eighties and nineties, there was a renaissance in this area of research with many proposed reconfigurable architectures developed in industry and academia, such as: Matrix, Garp,[3] Elixent, XPP, Silicon Hive, Montium, Pleiades, Morphosys, PiCoGA. Such designs were feasible due to the constant progress of silicon technology that let complex designs be implemented on one chip. The world's first commercial reconfigurable computer, the Algotronix CHS2X4, was completed in 1991. It was not a commercial success, but was promising enough that Xilinx (the inventor of the Field-Programmable Gate Array, FPGA) bought the technology and hired the Algotronix staff.[4]
Current systems
Currently there are a number of vendors with commercially available reconfigurable computers aimed at the high performance computing market; including Cray, SGI and SRC Computers, Inc. . The reconfigurable computers are "Estrin" hybrid computers with microprocessors that can be used in traditional CPU cluster computers or coupled to user-programmable FPGAs for hybrid computing. Cray supercomputer company (not affiliated with SRC Computers) acquired OctigaBay and its reconfigurable computing platform, which Cray marketed as the XD1 until recently. SGI sells the RASC platform with their Altix series of supercomputers.[5] SRC Computers, Inc. has developed a family of reconfigurable computers based on their IMPLICIT+EXPLICIT architecture and MAP processor.
The XD1 and SGI FPGA reconfiguration can be accomplished either via the traditional Hardware Description Languages (HDL), which can be generated directly or by using electronic design automation (“EDA”) or electronic system level (“ESL”) tools, employing high level languages like the graphical tool Starbridge Viva or C-based languages like for example Handel-C from Celoxica, DIME-C from Nallatech, Impulse-C from Impulse Accelerated Technologies or Mitrion-C from Mitrionics.
In addition, Mitrionics has developed a Software Acceleration Platform that enables software written using a single assignment language to be compiled and executed on FPGA-based computers, such as those from Cray and SGI. The Mitrion-C software language and Mitrion Virtual Processor enable software developers to write and execute applications on FPGA-based computers in the same manner as with other computing technologies, such as graphical processing units (“GPUs”), cell-based processors, parallel processing units (“PPUs”), multi-core CPUs, and traditional single-core CPU clusters.
SRC has developed a "Carte" compiler that takes an existing high-level languages like C or Fortran, and with a few modifications, compiles them for execution on both the FPGA and microprocessor. According to SRC literature, "...application algorithms are written in a high-level language such as C or Fortran. Carte extracts the maximum parallelism from the code and generates pipelined hardware logic that is instantiated in the MAP. It also generates all the required interface code to manage the movement of data to and from the MAP and to coordinate the microprocessor with the logic running in the MAP." (note that SRC also allows a traditional HDL flow to be used). The SRC systems communicate via the SNAP memory interface, and/or the (optional) Hi-Bar switch.
Research community is also acting on the subject with projects like MORPHEUS [6] in Europe which implements on a single 100mm² 90nm chip an ARM9 processor, a M2000 FPGA, a DREAM picoGA and an XPP Pact matrix.
Comparison of systems
As an emerging field, classifications of reconfigurable architectures are still being developed and refined as new architectures are developed; no unifying taxonomy has been suggested to date. However, several recurring parameters can be used to classify these systems.
Granularity
The granularity of the reconfigurable logic is defined as the size of the smallest functional unit (CLB) that is addressed by the mapping tools. Low granularity, which can also be known as fine-grained, often implies a greater flexibility when implementing algorithms into the hardware. However, there is a penalty associated with this in terms of increased power, area and delay due to greater quantity of routing required per computation. Fine-grained architectures work at the bit-level manipulation level; whilst coarse grained processing elements (rDPU) are better optimised for standard data path applications. One of the drawbacks of coarse grained architectures are that they tend to lose some of their utilisation and performance if they need to perform smaller computations than their granularity provides, for example for a one bit add on a four bit wide functional unit would waste three bits. This problem can be solved by having a coarse grain array (rDPA) and a FPGA on the same chip.
Coarse-grained architectures (rDPA) are intended for the implementation for algorithms needing word-width data paths (rDPU). As their functional blocks are optimized for large computations they will perform these operations more quickly and power efficiently than a smaller set of functional units connected together with some interconnect, this is due to the connecting wires are shorter, meaning less wire capacitance and hence faster and lower power designs. A potential undesirable consequence of having larger computational blocks is that when the size of operands may not match the algorithm an inefficient utilisation of resources can result. Often the type of applications to be run are known in advance allowing the logic, memory and routing resources to be tailored (for instance, see KressArray Xplorer) to enhance the performance of the device whilst still providing a certain level of flexibility for future adaptation. Examples of this are domain specific arrays aimed at gaining better performance in terms of power, area, throughput than their more generic finer grained FPGA cousins by reducing their flexibility.
Rate of reconfiguration
Configuration of these reconfigurable systems can happen at deployment time, between execution phases or during execution. In a typical reconfigurable system, a bit stream is used to program the device at deployment time. Fine grained systems by their own nature requires greater configuration time than more coarse-grained architectures due to more elements needing to be addressed and programmed. Therefore more coarse-grained architectures gain from potential lower energy requirements, as less information is transferred and utilised. Intuitively, the slower the rate of reconfiguration the smaller the energy consumption as the associated energy cost of reconfiguration are amortised over a longer period of time. Partial reconfiguration aims to allow part of the device to be reprogrammed while another part is still performing active computation. Partial reconfiguration allows smaller reconfigurable bit streams thus not wasting energy on transmitting redundant information in the bit stream. Compression of the bit stream is possible but careful analysis is to be carried out to insure that the energy saved by using smaller bit streams is not outweighed by the computation needed to decompress the data.
Host coupling
Often the reconfigurable array is used as a processing accelerator attached to a host processor. The level of coupling determines the type of data transfers, latency, power, throughput and overheads involved when utilising the reconfigurable logic. Some of the most intuitive designs use a peripheral bus to provide a coprocessor like arrangement for the reconfigurable array. However, there have also been implementations where the reconfigurable fabric is much closer to the processor, some are even implemented into the data path, utilising the processor registers. The job of the host processor is to perform the control functions, configure the logic, schedule data and to provide external interfacing.
Routing/interconnects
The flexibility in reconfigurable devices mainly comes from their routing interconnect. One style of interconnect made popular by FPGAs vendors, Xilinx and Altera are the island style layout, where blocks are arranged in an array with vertical and horizontal routing. A layout with inadequate routing may suffer from poor flexibility and resource utilisation, therefore providing limited performance. If too much interconnect is provided this requires more transistors than necessary and thus more silicon area, longer wires and more power consumption.
Tool flow
Generally, tools for configurable computing systems can be split up in two parts, CAD tools for reconfigurable array and compilation tools for CPU. The front-end compiler is an integrated tool, and will generate a structural hardware representation that is input of hardware design flow. Hardware design flow for reconfigurable architecture can be classified by the approach adopted by three main stages of design process: technology mapping, placement algorithm and routing algorithm. The software frameworks differ in the level of the programming language.
Some types of reconfigurable computers are microcoded processors where the microcode is stored in RAM or EEPROM, and changeable on reboot or on the fly. This could be done with the AMD 2900 series bit slice processors (on reboot) and later with FPGAs (on the fly).
Some dataflow processors are implemented using reconfigurable computing.
To compare the effect of various ways to implement an algorithm on the runtime and energy used, some tools allow compiling the same piece of C code for a fixed CPU, a soft processor, or compiling directly to FPGA [2].
Reconfigurable computing as a paradigm shift: Hartenstein's anti machine
Table 1: Nick Tredennick’s Paradigm Classification Scheme Early Historic Computers:
Programming Source
Resources fixed none
Algorithms fixed none
von Neumann Computer:
Programming Source
Resources fixed none
Algorithms variable Software (instruction streams)
Reconfigurable Computing Systems:
Programming Source
Resources variable Configware (configuration)
Algorithms variable Flowware (data streams)
Computer scientist Reiner Hartenstein describes reconfigurable computing in terms of an anti machine that, according to him, represents a fundamental paradigm shift away from the more conventional von Neumann machine .[7] Hartenstein describes a Reconfigurable Computing Paradox: [8] Software to configware migration (software to FPGA migration) results in reported speed-up factors of up to almost four orders of magnitude, as well as a reduction in electricity consumption by more than one order of magnitude---although the technological parameters of FPGA's are behind the Gordon Moore curve by about four orders of magnitude, and the clock frequency is substantially lower than that of microprocessors. This paradox is due to a paradigm shift, and is also partly explained by the Von Neumann syndrome.
The fundamental model of the reconfigurable computing machine paradigm, the data-stream-based anti machine is well illustrated by the differences to other machine paradigms that were introduced earlier, as shown by Nick Tredennick's following classification scheme of computing paradigms (see "Table 1: Nick Tredennick’s Paradigm Classification Scheme") .[9]
The fundamental model of a Reconfigurable Computing Machine, the data-stream-based anti machine (also called Xputer), is the counterpart of the instruction-stream-based von Neumann machine paradigm. This is illustrated by a simple reconfigurable system (not dynamically reconfigurable), which has no instruction fetch at run time. The reconfiguration (before run time) can be considered as a kind of super instruction fetch. An anti machine does not have a program counter. The anti machine has data counters instead, since it is data-stream-driven. Here the definition of the term data streams is adopted from the systolic array scene, which defines, at which time which data item has to enter or leave which port, here of the reconfigurable system, which may be fine-grained (e. g. using FPGAs) or coarse-grained, or a mixture of both.
The systolic array scene, originally (early 80ies) mainly mathematicians, only defined one half of the anti machine: the data path: the Systolic array (also see Super Systolic Array). But they did not define nor model the data sequencer methodology, considering that this is not their job to take care where the data streams come from or end up. The data sequencing part of the anti machine is modeled as distributed memory, preferrably on chip, which consists of auto-sequencing memory blocks (ASM blocks). Each ASM block has a sequencer including a data counter. An example is the Generic Address Generator (GAG), which is a generalization of the DMA.
The concept of reconfigurable computing has existed since the 1960s, when Gerald Estrin's landmark paper proposed the concept of a computer made of a standard processor and an array of “reconfigurable” hardware.[1][2] The main processor would control the behavior of the reconfigurable hardware. The latter would then be tailored to perform a specific task, such as image processing or pattern matching, as quickly as a dedicated piece of hardware. Once the task was done, the hardware could be adjusted to do some other task. This resulted in a hybrid computer structure combining the flexibility of software with the speed of hardware; unfortunately this idea was far ahead of its time in needed electronic technology.
In the eighties and nineties, there was a renaissance in this area of research with many proposed reconfigurable architectures developed in industry and academia, such as: Matrix, Garp,[3] Elixent, XPP, Silicon Hive, Montium, Pleiades, Morphosys, PiCoGA. Such designs were feasible due to the constant progress of silicon technology that let complex designs be implemented on one chip. The world's first commercial reconfigurable computer, the Algotronix CHS2X4, was completed in 1991. It was not a commercial success, but was promising enough that Xilinx (the inventor of the Field-Programmable Gate Array, FPGA) bought the technology and hired the Algotronix staff.[4]
Current systems
Currently there are a number of vendors with commercially available reconfigurable computers aimed at the high performance computing market; including Cray, SGI and SRC Computers, Inc. . The reconfigurable computers are "Estrin" hybrid computers with microprocessors that can be used in traditional CPU cluster computers or coupled to user-programmable FPGAs for hybrid computing. Cray supercomputer company (not affiliated with SRC Computers) acquired OctigaBay and its reconfigurable computing platform, which Cray marketed as the XD1 until recently. SGI sells the RASC platform with their Altix series of supercomputers.[5] SRC Computers, Inc. has developed a family of reconfigurable computers based on their IMPLICIT+EXPLICIT architecture and MAP processor.
The XD1 and SGI FPGA reconfiguration can be accomplished either via the traditional Hardware Description Languages (HDL), which can be generated directly or by using electronic design automation (“EDA”) or electronic system level (“ESL”) tools, employing high level languages like the graphical tool Starbridge Viva or C-based languages like for example Handel-C from Celoxica, DIME-C from Nallatech, Impulse-C from Impulse Accelerated Technologies or Mitrion-C from Mitrionics.
In addition, Mitrionics has developed a Software Acceleration Platform that enables software written using a single assignment language to be compiled and executed on FPGA-based computers, such as those from Cray and SGI. The Mitrion-C software language and Mitrion Virtual Processor enable software developers to write and execute applications on FPGA-based computers in the same manner as with other computing technologies, such as graphical processing units (“GPUs”), cell-based processors, parallel processing units (“PPUs”), multi-core CPUs, and traditional single-core CPU clusters.
SRC has developed a "Carte" compiler that takes an existing high-level languages like C or Fortran, and with a few modifications, compiles them for execution on both the FPGA and microprocessor. According to SRC literature, "...application algorithms are written in a high-level language such as C or Fortran. Carte extracts the maximum parallelism from the code and generates pipelined hardware logic that is instantiated in the MAP. It also generates all the required interface code to manage the movement of data to and from the MAP and to coordinate the microprocessor with the logic running in the MAP." (note that SRC also allows a traditional HDL flow to be used). The SRC systems communicate via the SNAP memory interface, and/or the (optional) Hi-Bar switch.
Research community is also acting on the subject with projects like MORPHEUS [6] in Europe which implements on a single 100mm² 90nm chip an ARM9 processor, a M2000 FPGA, a DREAM picoGA and an XPP Pact matrix.
Comparison of systems
As an emerging field, classifications of reconfigurable architectures are still being developed and refined as new architectures are developed; no unifying taxonomy has been suggested to date. However, several recurring parameters can be used to classify these systems.
Granularity
The granularity of the reconfigurable logic is defined as the size of the smallest functional unit (CLB) that is addressed by the mapping tools. Low granularity, which can also be known as fine-grained, often implies a greater flexibility when implementing algorithms into the hardware. However, there is a penalty associated with this in terms of increased power, area and delay due to greater quantity of routing required per computation. Fine-grained architectures work at the bit-level manipulation level; whilst coarse grained processing elements (rDPU) are better optimised for standard data path applications. One of the drawbacks of coarse grained architectures are that they tend to lose some of their utilisation and performance if they need to perform smaller computations than their granularity provides, for example for a one bit add on a four bit wide functional unit would waste three bits. This problem can be solved by having a coarse grain array (rDPA) and a FPGA on the same chip.
Coarse-grained architectures (rDPA) are intended for the implementation for algorithms needing word-width data paths (rDPU). As their functional blocks are optimized for large computations they will perform these operations more quickly and power efficiently than a smaller set of functional units connected together with some interconnect, this is due to the connecting wires are shorter, meaning less wire capacitance and hence faster and lower power designs. A potential undesirable consequence of having larger computational blocks is that when the size of operands may not match the algorithm an inefficient utilisation of resources can result. Often the type of applications to be run are known in advance allowing the logic, memory and routing resources to be tailored (for instance, see KressArray Xplorer) to enhance the performance of the device whilst still providing a certain level of flexibility for future adaptation. Examples of this are domain specific arrays aimed at gaining better performance in terms of power, area, throughput than their more generic finer grained FPGA cousins by reducing their flexibility.
Rate of reconfiguration
Configuration of these reconfigurable systems can happen at deployment time, between execution phases or during execution. In a typical reconfigurable system, a bit stream is used to program the device at deployment time. Fine grained systems by their own nature requires greater configuration time than more coarse-grained architectures due to more elements needing to be addressed and programmed. Therefore more coarse-grained architectures gain from potential lower energy requirements, as less information is transferred and utilised. Intuitively, the slower the rate of reconfiguration the smaller the energy consumption as the associated energy cost of reconfiguration are amortised over a longer period of time. Partial reconfiguration aims to allow part of the device to be reprogrammed while another part is still performing active computation. Partial reconfiguration allows smaller reconfigurable bit streams thus not wasting energy on transmitting redundant information in the bit stream. Compression of the bit stream is possible but careful analysis is to be carried out to insure that the energy saved by using smaller bit streams is not outweighed by the computation needed to decompress the data.
Host coupling
Often the reconfigurable array is used as a processing accelerator attached to a host processor. The level of coupling determines the type of data transfers, latency, power, throughput and overheads involved when utilising the reconfigurable logic. Some of the most intuitive designs use a peripheral bus to provide a coprocessor like arrangement for the reconfigurable array. However, there have also been implementations where the reconfigurable fabric is much closer to the processor, some are even implemented into the data path, utilising the processor registers. The job of the host processor is to perform the control functions, configure the logic, schedule data and to provide external interfacing.
Routing/interconnects
The flexibility in reconfigurable devices mainly comes from their routing interconnect. One style of interconnect made popular by FPGAs vendors, Xilinx and Altera are the island style layout, where blocks are arranged in an array with vertical and horizontal routing. A layout with inadequate routing may suffer from poor flexibility and resource utilisation, therefore providing limited performance. If too much interconnect is provided this requires more transistors than necessary and thus more silicon area, longer wires and more power consumption.
Tool flow
Generally, tools for configurable computing systems can be split up in two parts, CAD tools for reconfigurable array and compilation tools for CPU. The front-end compiler is an integrated tool, and will generate a structural hardware representation that is input of hardware design flow. Hardware design flow for reconfigurable architecture can be classified by the approach adopted by three main stages of design process: technology mapping, placement algorithm and routing algorithm. The software frameworks differ in the level of the programming language.
Some types of reconfigurable computers are microcoded processors where the microcode is stored in RAM or EEPROM, and changeable on reboot or on the fly. This could be done with the AMD 2900 series bit slice processors (on reboot) and later with FPGAs (on the fly).
Some dataflow processors are implemented using reconfigurable computing.
To compare the effect of various ways to implement an algorithm on the runtime and energy used, some tools allow compiling the same piece of C code for a fixed CPU, a soft processor, or compiling directly to FPGA [2].
Reconfigurable computing as a paradigm shift: Hartenstein's anti machine
Table 1: Nick Tredennick’s Paradigm Classification Scheme Early Historic Computers:
Programming Source
Resources fixed none
Algorithms fixed none
von Neumann Computer:
Programming Source
Resources fixed none
Algorithms variable Software (instruction streams)
Reconfigurable Computing Systems:
Programming Source
Resources variable Configware (configuration)
Algorithms variable Flowware (data streams)
Computer scientist Reiner Hartenstein describes reconfigurable computing in terms of an anti machine that, according to him, represents a fundamental paradigm shift away from the more conventional von Neumann machine .[7] Hartenstein describes a Reconfigurable Computing Paradox: [8] Software to configware migration (software to FPGA migration) results in reported speed-up factors of up to almost four orders of magnitude, as well as a reduction in electricity consumption by more than one order of magnitude---although the technological parameters of FPGA's are behind the Gordon Moore curve by about four orders of magnitude, and the clock frequency is substantially lower than that of microprocessors. This paradox is due to a paradigm shift, and is also partly explained by the Von Neumann syndrome.
The fundamental model of the reconfigurable computing machine paradigm, the data-stream-based anti machine is well illustrated by the differences to other machine paradigms that were introduced earlier, as shown by Nick Tredennick's following classification scheme of computing paradigms (see "Table 1: Nick Tredennick’s Paradigm Classification Scheme") .[9]
The fundamental model of a Reconfigurable Computing Machine, the data-stream-based anti machine (also called Xputer), is the counterpart of the instruction-stream-based von Neumann machine paradigm. This is illustrated by a simple reconfigurable system (not dynamically reconfigurable), which has no instruction fetch at run time. The reconfiguration (before run time) can be considered as a kind of super instruction fetch. An anti machine does not have a program counter. The anti machine has data counters instead, since it is data-stream-driven. Here the definition of the term data streams is adopted from the systolic array scene, which defines, at which time which data item has to enter or leave which port, here of the reconfigurable system, which may be fine-grained (e. g. using FPGAs) or coarse-grained, or a mixture of both.
The systolic array scene, originally (early 80ies) mainly mathematicians, only defined one half of the anti machine: the data path: the Systolic array (also see Super Systolic Array). But they did not define nor model the data sequencer methodology, considering that this is not their job to take care where the data streams come from or end up. The data sequencing part of the anti machine is modeled as distributed memory, preferrably on chip, which consists of auto-sequencing memory blocks (ASM blocks). Each ASM block has a sequencer including a data counter. An example is the Generic Address Generator (GAG), which is a generalization of the DMA.
Reconfigurable computing
Reconfigurable computing is a computing paradigm combining some of the flexibility of software with the high performance of hardware by processing with very flexible high speed computing fabrics like FPGAs. The principal difference when compared to using ordinary microprocessors is the ability to make substantial changes to the data path itself in addition to the control flow. On the other hand, the main difference with custom hardware (ASICs) is the possibility to adapt the hardware during runtime by "loading" a new circuit on the reconfigurable fabric.
Quantum decoherence
One major problem is keeping the components of the computer in a coherent state, as the slightest interaction with the external world would cause the system to decohere. This effect causes the unitary character (and more specifically, the invertibility) of quantum computational steps to be violated. Decoherence times for candidate systems, in particular the transverse relaxation time T2 (terminology used in NMR and MRI technology, also called the dephasing time), typically range between nanoseconds and seconds at low temperature.[8] The issue for optical approaches are more difficult as these timescales are orders of magnitude lower and an often cited approach to overcome it uses an optical pulse shaping approach. Error rates are typically proportional to the ratio of operating time to decoherence time, hence any operation must be completed much more quickly than the decoherence time.
If the error rate is small enough, it is thought to be possible to use quantum error correction, which corrects errors due to decoherence, thereby allowing the total calculation time to be longer than the decoherence time. An often cited (but rather arbitrary) figure for required error rate in each gate is 10−4. This implies that each gate must be able to perform its task 10,000 times faster than the decoherence time of the system.
Meeting this scalability condition is possible for a wide range of systems. However, the use of error correction brings with it the cost of a greatly increased number of required qubits. The number required to factor integers using Shor's algorithm is still polynomial, and thought to be between L and L2, where L is the number of bits in the number to be factored. For a 1000-bit number, this implies a need for c103 to 106 qubits.[9] With error correction need about 1000 times more qubits.
A very different approach to the stability-decoherence problem is to create a topological quantum computer with anyons, quasi-particles used as threads and relying on knot theory to form stable logic gates
If the error rate is small enough, it is thought to be possible to use quantum error correction, which corrects errors due to decoherence, thereby allowing the total calculation time to be longer than the decoherence time. An often cited (but rather arbitrary) figure for required error rate in each gate is 10−4. This implies that each gate must be able to perform its task 10,000 times faster than the decoherence time of the system.
Meeting this scalability condition is possible for a wide range of systems. However, the use of error correction brings with it the cost of a greatly increased number of required qubits. The number required to factor integers using Shor's algorithm is still polynomial, and thought to be between L and L2, where L is the number of bits in the number to be factored. For a 1000-bit number, this implies a need for c103 to 106 qubits.[9] With error correction need about 1000 times more qubits.
A very different approach to the stability-decoherence problem is to create a topological quantum computer with anyons, quasi-particles used as threads and relying on knot theory to form stable logic gates
The power of quantum computers
Integer factorization is believed to be computationally infeasible with an ordinary computer for large integers that are the product of only a few prime numbers (e.g., products of two 300-digit primes).[6] By comparison, a quantum computer could solve this problem more efficiently than a classical computer using Shor's algorithm to find its factors. This ability would allow a quantum computer to "break" many of the cryptographic systems in use today, in the sense that there would be a polynomial time (in the number of bits of the integer) algorithm for solving the problem. In particular, most of the popular public key ciphers are based on the difficulty of factoring integers, including forms of RSA. These are used to protect secure Web pages, encrypted email, and many other types of data. Breaking these would have significant ramifications for electronic privacy and security. The only way to increase the security of an algorithm like RSA would be to increase the key size and hope that an adversary does not have the resources to build and use a powerful enough quantum computer.
A way out of this dilemma would be to use some kind of quantum cryptography. There are also some digital signature schemes that are believed to be secure against quantum computers. See for instance Lamport signatures.
This dramatic advantage of quantum computers has only been discovered for these problems so far: factoring, discrete logarithm. However, there is no proof that the advantage is real: an equally fast classical algorithm may still be discovered. There is one other problem where quantum computers have a smaller, though significant (quadratic) advantage. It is quantum database search, and can be solved by Grover's algorithm. In this case the advantage is provable. This establishes beyond doubt that (ideal) quantum computers are superior to classical computers for at least one problem.
Consider a problem that has these four properties:
1. The only way to solve it is to guess answers repeatedly and check them,
2. There are n possible answers to check,
3. Every possible answer takes the same amount of time to check, and
4. There are no clues about which answers might be better: generating possibilities randomly is just as good as checking them in some special order.
An example of this is a password cracker that attempts to guess the password for an encrypted file (assuming that the password has a maximum possible length).
For problems with all four properties, the time for a quantum computer to solve this will be proportional to the square root of n (it would take an average of (n + 1)/2 guesses to find the answer using a classical computer.) That can be a very large speedup, reducing some problems from years to seconds. It can be used to attack symmetric ciphers such as Triple DES and AES by attempting to guess the secret key. Regardless of whether any of these problems can be shown to have an advantage on a quantum computer, they nonetheless will always have the advantage of being an excellent tool for studying quantum mechanical interactions, which of itself is an enormous value to the scientific community.
Grover's algorithm can also be used to obtain a quadratic speed-up for a class of problems known as NP-complete.
A way out of this dilemma would be to use some kind of quantum cryptography. There are also some digital signature schemes that are believed to be secure against quantum computers. See for instance Lamport signatures.
This dramatic advantage of quantum computers has only been discovered for these problems so far: factoring, discrete logarithm. However, there is no proof that the advantage is real: an equally fast classical algorithm may still be discovered. There is one other problem where quantum computers have a smaller, though significant (quadratic) advantage. It is quantum database search, and can be solved by Grover's algorithm. In this case the advantage is provable. This establishes beyond doubt that (ideal) quantum computers are superior to classical computers for at least one problem.
Consider a problem that has these four properties:
1. The only way to solve it is to guess answers repeatedly and check them,
2. There are n possible answers to check,
3. Every possible answer takes the same amount of time to check, and
4. There are no clues about which answers might be better: generating possibilities randomly is just as good as checking them in some special order.
An example of this is a password cracker that attempts to guess the password for an encrypted file (assuming that the password has a maximum possible length).
For problems with all four properties, the time for a quantum computer to solve this will be proportional to the square root of n (it would take an average of (n + 1)/2 guesses to find the answer using a classical computer.) That can be a very large speedup, reducing some problems from years to seconds. It can be used to attack symmetric ciphers such as Triple DES and AES by attempting to guess the secret key. Regardless of whether any of these problems can be shown to have an advantage on a quantum computer, they nonetheless will always have the advantage of being an excellent tool for studying quantum mechanical interactions, which of itself is an enormous value to the scientific community.
Grover's algorithm can also be used to obtain a quadratic speed-up for a class of problems known as NP-complete.
Initialization, execution and termination
In our example, the contents of the qubit registers can be thought of as an 8-dimensional complex vector. An algorithm for a quantum computer must initialize this vector in some specified form (dependent on the design of the quantum computer). In each step of the algorithm, that vector is modified by multiplying it by a unitary matrix. The matrix is determined by the physics of the device. The unitary character of the matrix ensures the matrix is invertible (so each step is reversible).
Upon termination of the algorithm, the 8-dimensional complex vector stored in the register must be somehow read off from the qubit register by a quantum measurement. However, by the laws of quantum mechanics, that measurement will yield a random 3-bit string (and it will destroy the stored state as well). This random string can be used in computing the value of a function because (by design) the probability distribution of the measured output bitstring is skewed in favor of the correct value of the function. By repeated runs of the quantum computer and measurement of the output, the correct value can be determined, to a high probability, by majority polling of the outputs. In brief, quantum computations are probabilistic; see quantum circuit for a more precise formulation.
For more details on the sequences of operations used for various algorithms, see universal quantum computer, Shor's algorithm, Grover's algorithm, Deutsch-Jozsa algorithm, quantum Fourier transform, quantum gate, quantum adiabatic algorithm and quantum error correction. Also refer to the growing field of quantum programming.
Upon termination of the algorithm, the 8-dimensional complex vector stored in the register must be somehow read off from the qubit register by a quantum measurement. However, by the laws of quantum mechanics, that measurement will yield a random 3-bit string (and it will destroy the stored state as well). This random string can be used in computing the value of a function because (by design) the probability distribution of the measured output bitstring is skewed in favor of the correct value of the function. By repeated runs of the quantum computer and measurement of the output, the correct value can be determined, to a high probability, by majority polling of the outputs. In brief, quantum computations are probabilistic; see quantum circuit for a more precise formulation.
For more details on the sequences of operations used for various algorithms, see universal quantum computer, Shor's algorithm, Grover's algorithm, Deutsch-Jozsa algorithm, quantum Fourier transform, quantum gate, quantum adiabatic algorithm and quantum error correction. Also refer to the growing field of quantum programming.
Bits vs. Qubits
Consider first a classical computer that operates on a 3-bit register. At any given time, the bits in the register are in a definite state, such as 101. In a quantum computer, however, the qubits can be in a superposition of all the classically allowed states. In fact, the register is described by a wavefunction:
|\psi \rangle = a\,|000\rangle + b\,|001\rangle + c\,|010\rangle + d\,|011\rangle + e\,|100\rangle + f\,|101\rangle + g\,|110\rangle + h\,|111\rangle
Qubits are made up of controlled particles and the means of control (e.g. devices that trap particles and switch them from one state to another).
Qubits are made up of controlled particles and the means of control (e.g. devices that trap particles and switch them from one state to another).[3]
where the coefficients a, b, c,..., h are complex numbers whose amplitudes squared are the probabilities to measure the qubits in each state- for example, |c|^2\, is the probability to measure the register in the state 010. It is important that these numbers are complex, because the phases of the numbers can constructively and destructively interfere with one another; this is an important feature for quantum algorithms.[4] The basis made up of 0 and 1 (true and false) is called the computational basis, but other bases can be used. Another common basis, used for example in measurement based quantum computation is the Hadamard basis of |+\rangle and |-\rangle. Any two orthogonal vectors can be used as a basis.
Recording the state of a quantum register requires an exponential number of complex numbers (the 3-qubit register above requires 23 = 8 complex numbers). The number of classical bits required even to estimate the complex numbers of some quantum state grows exponentially with the number of qubits. For a 300-qubit quantum register, somewhere on the order of 2300 = 1090 classical registers are required, more than there are atoms in the observable universe.
|\psi \rangle = a\,|000\rangle + b\,|001\rangle + c\,|010\rangle + d\,|011\rangle + e\,|100\rangle + f\,|101\rangle + g\,|110\rangle + h\,|111\rangle
Qubits are made up of controlled particles and the means of control (e.g. devices that trap particles and switch them from one state to another).
Qubits are made up of controlled particles and the means of control (e.g. devices that trap particles and switch them from one state to another).[3]
where the coefficients a, b, c,..., h are complex numbers whose amplitudes squared are the probabilities to measure the qubits in each state- for example, |c|^2\, is the probability to measure the register in the state 010. It is important that these numbers are complex, because the phases of the numbers can constructively and destructively interfere with one another; this is an important feature for quantum algorithms.[4] The basis made up of 0 and 1 (true and false) is called the computational basis, but other bases can be used. Another common basis, used for example in measurement based quantum computation is the Hadamard basis of |+\rangle and |-\rangle. Any two orthogonal vectors can be used as a basis.
Recording the state of a quantum register requires an exponential number of complex numbers (the 3-qubit register above requires 23 = 8 complex numbers). The number of classical bits required even to estimate the complex numbers of some quantum state grows exponentially with the number of qubits. For a 300-qubit quantum register, somewhere on the order of 2300 = 1090 classical registers are required, more than there are atoms in the observable universe.
The basis of quantum computing
A classical computer has a memory made up of bits, where each bit holds either a one or a zero. A quantum computer maintains a sequence of qubits. A single qubit can hold a one, a zero, or, crucially, a quantum superposition of these, and any two qubits can be in a quantum superposition of 4 states, and three qubits in 8. In general a quantum computer with n qubits can be in up to 2n different states simultaneously (this compares to a normal computer that can only be in one of 2n states at any one time). A quantum computer operates by manipulating those qubits with (possibly a suite of) quantum logic gates.
An example of an implementation of qubits for a quantum computer could start with the use of particles with two spin states: "up" and "down" (typically written |0\rangle and |1\rangle). But in fact any system possessing an observable quantity A which is conserved under time evolution and such that A has at least two discrete and sufficiently spaced consecutive eigenvalues, is a suitable candidate for implementing a qubit. This is true because any such system can be mapped onto an effective spin-1/2.
An example of an implementation of qubits for a quantum computer could start with the use of particles with two spin states: "up" and "down" (typically written |0\rangle and |1\rangle). But in fact any system possessing an observable quantity A which is conserved under time evolution and such that A has at least two discrete and sufficiently spaced consecutive eigenvalues, is a suitable candidate for implementing a qubit. This is true because any such system can be mapped onto an effective spin-1/2.
Quantum computer
A quantum computer is a device for computation that makes direct use of distinctively quantum mechanical phenomena, such as superposition and entanglement, to perform operations on data. In a classical (or conventional) computer, information is stored as bits; in a quantum computer, it is stored as qubits (quantum bits). The basic principle of quantum computation is that the quantum properties can be used to represent and structure data, and that quantum mechanisms can be devised and built to perform operations with this data.[1]
Although quantum computing is still in its infancy, experiments have been carried out in which quantum computational operations were executed on a very small number of qubits. Research in both theoretical and practical areas continues at a frantic pace, and many national government and military funding agencies support quantum computing research to develop quantum computers for both civilian and national security purposes, such as cryptanalysis.[2] (See Timeline of quantum computing for details on current and past progress.)
If large-scale quantum computers can be built, they will be able to solve certain problems much faster than any of our current classical computers (for example Shor's algorithm). Quantum computers are different from other computers such as DNA computers and traditional computers based on transistors. Some computing architectures such as optical computers may use classical superposition of electromagnetic waves, but without some specifically quantum mechanical resources such as entanglement, they do not have the same computational speed-up as quantum computers.
Although quantum computing is still in its infancy, experiments have been carried out in which quantum computational operations were executed on a very small number of qubits. Research in both theoretical and practical areas continues at a frantic pace, and many national government and military funding agencies support quantum computing research to develop quantum computers for both civilian and national security purposes, such as cryptanalysis.[2] (See Timeline of quantum computing for details on current and past progress.)
If large-scale quantum computers can be built, they will be able to solve certain problems much faster than any of our current classical computers (for example Shor's algorithm). Quantum computers are different from other computers such as DNA computers and traditional computers based on transistors. Some computing architectures such as optical computers may use classical superposition of electromagnetic waves, but without some specifically quantum mechanical resources such as entanglement, they do not have the same computational speed-up as quantum computers.
Redshift Theory
Redshift is a techno-economic theory suggesting hypersegmentation of Information Technology (IT) markets based on whether individual computing needs are over or under-served by Moore's Law, which predicts the doubling of computing transistors (and therefore roughly computing power) every two years. The theory, proposed and named by Sun Microsystems CTO Greg Papadopoulos, categorizes a series of high growth markets (Redshifting) while predicting slower GDP-driven growth in traditional computing markets (Blueshifting). Papadopoulos predicts the result will be a fundamental redesign of components comprising computing systems.
Hypergrowth Market Segments (Redshifting)
According to the Redshift theory, “Redshifting” applications are growing dramatically faster than Moore's Law, growing quickly in their absolute number of systems.In these markets, customers are running out of datacenter real estate, power and cooling infrastructure.According to Dell Senior Vice President Brad Anderson, “Businesses requiring hyper-scale computing environments – where infrastructure deployments are measured by up to millions of servers, storage and networking equipment – are changing the way they approach IT.”
While various Redshift proponents offer minor alterations on the original presentation, “Redshifting” generally includes:
ΣBW (Sum-of-Bandwidth)
Companies who are driving heavy Internet traffic. This includes popular web portals like Google, Yahoo, AOL and MSN. It also includes telecoms, multimedia, television over IP, online games like World of Warcraft and others. This segment has been enabled by widespread availability of high bandwidth Internet connections to consumers through a DSL or cable modem. A simple way to understand this market is that for every byte of content served to a PC, mobile phone or other device over a network, there must exist computing systems to send it over the network.
High Performance Computing (HPC)
Companies doing complex simulations such as weather, stock market or drug design simulations. This is a generally elastic market because businesses frequently spend every "available" dollar budgeted for IT. A common anecdote is that cutting the cost of computing by half causes customers in this segment to buy at least twice as much, because each marginal IT dollar spent contributes to business advantage.
*prise (or "Star-prise")
Companies aggregating traditional computing applications and offer them as services, typically in the form of Software as a Service (SaaS). For example, companies deploying CRM solutions are over-served by Moore's Law, but companies that aggregate CRM solutions and offer them as a service, such as Salesforce.com, are growing faster than Moore's Law.
Traditional Computing Markets (Blueshifting)
Redshift theory suggests that traditional computing markets, such as those serving Enterprise Resource Planning (ERP) or Customer Relationship Management (CRM) applications, have reached relative saturation in industrialized nations. Thereafter, proponents argue further market growth will closely follow Gross Domestic Product (GDP) growth, which typically remains under 10% for most countries annually. Given that Moore's Law continues to predict accurately the rate of computing transistor growth, which roughly translates into computing power doubling every two years, the Redshift theory suggests that traditional computing markets will ultimately contract as a percentage of computing expenditures over time.
Functionally, this means “Blueshifting” customers can satisfy computing requirement growth by swapping in faster processors without increasing the absolute number of computing systems.
Consequences & Industry Commentary
Papadopoulos argues that while traditional computing markets remain the dominant source of revenue through the late 2000s, a shift to hypergrowth markets will inevitably occur. When that shift occurs, he argues computing (but not computers) will become a utility, and differentiation in the IT market will be based upon a company's ability to deliver computing at massive scale, efficiently and with predictable service levels, much like electricity today.
If computing is to be delivered as a utility, Nicholas Carr suggests Papadopoulos' vision compares with Microsoft researcher Jim Hamilton, who both agree that computing is most efficiently generated in shipping containers.[4] Industry analysts are also beginning to quantify Redshifting and Blueshifting markets. According to International Data Corporation (IDC) vice president Matthew Eastwood, "IDC believes that the IT market is in a period of hyper segmentation... This a class of customers that is Moore's law driven and as price performance gains continue, IDC believes that these organizations will accelerate their consumption of IT infrastructure.”
History & Nomenclature
Key portions of Papadopoulos' theory were first presented by Sun Microsystems CEO Jonathan Schwartz in late 2006.[6] Papadopoulos later gave a full presentation on Redshift to Sun's annual Analyst Summit[1] in February 2007. The term Redshift refers to what happens when electromagnetic radiation, usually visible light, moves away from an observer. Papadopoulos chose this term to reflect growth markets because redshift helped cosmologists explain the expansion of the universe.
Papadopoulos originally depicted traditional IT markets as green to represent their revenue base, but later changed them to “blueshift,” which occurs when a light source moves toward an observer, similar to what would happen during a contraction of the universe.
Hypergrowth Market Segments (Redshifting)
According to the Redshift theory, “Redshifting” applications are growing dramatically faster than Moore's Law, growing quickly in their absolute number of systems.In these markets, customers are running out of datacenter real estate, power and cooling infrastructure.According to Dell Senior Vice President Brad Anderson, “Businesses requiring hyper-scale computing environments – where infrastructure deployments are measured by up to millions of servers, storage and networking equipment – are changing the way they approach IT.”
While various Redshift proponents offer minor alterations on the original presentation, “Redshifting” generally includes:
ΣBW (Sum-of-Bandwidth)
Companies who are driving heavy Internet traffic. This includes popular web portals like Google, Yahoo, AOL and MSN. It also includes telecoms, multimedia, television over IP, online games like World of Warcraft and others. This segment has been enabled by widespread availability of high bandwidth Internet connections to consumers through a DSL or cable modem. A simple way to understand this market is that for every byte of content served to a PC, mobile phone or other device over a network, there must exist computing systems to send it over the network.
High Performance Computing (HPC)
Companies doing complex simulations such as weather, stock market or drug design simulations. This is a generally elastic market because businesses frequently spend every "available" dollar budgeted for IT. A common anecdote is that cutting the cost of computing by half causes customers in this segment to buy at least twice as much, because each marginal IT dollar spent contributes to business advantage.
*prise (or "Star-prise")
Companies aggregating traditional computing applications and offer them as services, typically in the form of Software as a Service (SaaS). For example, companies deploying CRM solutions are over-served by Moore's Law, but companies that aggregate CRM solutions and offer them as a service, such as Salesforce.com, are growing faster than Moore's Law.
Traditional Computing Markets (Blueshifting)
Redshift theory suggests that traditional computing markets, such as those serving Enterprise Resource Planning (ERP) or Customer Relationship Management (CRM) applications, have reached relative saturation in industrialized nations. Thereafter, proponents argue further market growth will closely follow Gross Domestic Product (GDP) growth, which typically remains under 10% for most countries annually. Given that Moore's Law continues to predict accurately the rate of computing transistor growth, which roughly translates into computing power doubling every two years, the Redshift theory suggests that traditional computing markets will ultimately contract as a percentage of computing expenditures over time.
Functionally, this means “Blueshifting” customers can satisfy computing requirement growth by swapping in faster processors without increasing the absolute number of computing systems.
Consequences & Industry Commentary
Papadopoulos argues that while traditional computing markets remain the dominant source of revenue through the late 2000s, a shift to hypergrowth markets will inevitably occur. When that shift occurs, he argues computing (but not computers) will become a utility, and differentiation in the IT market will be based upon a company's ability to deliver computing at massive scale, efficiently and with predictable service levels, much like electricity today.
If computing is to be delivered as a utility, Nicholas Carr suggests Papadopoulos' vision compares with Microsoft researcher Jim Hamilton, who both agree that computing is most efficiently generated in shipping containers.[4] Industry analysts are also beginning to quantify Redshifting and Blueshifting markets. According to International Data Corporation (IDC) vice president Matthew Eastwood, "IDC believes that the IT market is in a period of hyper segmentation... This a class of customers that is Moore's law driven and as price performance gains continue, IDC believes that these organizations will accelerate their consumption of IT infrastructure.”
History & Nomenclature
Key portions of Papadopoulos' theory were first presented by Sun Microsystems CEO Jonathan Schwartz in late 2006.[6] Papadopoulos later gave a full presentation on Redshift to Sun's annual Analyst Summit[1] in February 2007. The term Redshift refers to what happens when electromagnetic radiation, usually visible light, moves away from an observer. Papadopoulos chose this term to reflect growth markets because redshift helped cosmologists explain the expansion of the universe.
Papadopoulos originally depicted traditional IT markets as green to represent their revenue base, but later changed them to “blueshift,” which occurs when a light source moves toward an observer, similar to what would happen during a contraction of the universe.
Methods For good Computing
Several methods are available for minimizing the environmental impact of a computing system.
Power generation
See also: Green energy
Computers can be powered by different sources, including windmills, hydroelectric, photovoltaic panels, or nuclear. Some of these sources are perceived to be more environmentally friendly than others. Other novel sources include human power, such as the Twibright Exciter project.
Virtualization
Main article: x86 virtualization
See also: Comparison of virtual machines
Computer virtualization is the process of running two or more logical computer systems on one set of physical hardware. The concept originated with the mainframe operating systems of the 1960s, but was commercialized for x86-compatible computers only in the 1990s. With virtualization, a system administrator could combine several physical systems into virtual machines on one single, powerful system, thereby unplugging the original hardware and reducing power and cooling consumption. Several commercial companies and open-source projects now offer software packages to enable a transition to virtual computing. Intel Corporation and AMD have also built proprietary virtualization enhancements to the x86 instruction set into each of their CPU product lines, in order to facilitate virtualized computing.
Power management
Main article: Power management
An open industry standard called Advanced Configuration and Power Interface (ACPI) provides a standard programming interface that allows an operating system to directly control the power saving aspects of the hardware. This allows the system to automatically turn off components such as monitors and hard drives after set periods of inactivity. In addition, a system may hibernate, in which it turns off nearly all components, including the CPU and the system RAM, greatly reducing the system's electricity usage. To resume from this state, some components, such as the keyboard, network interface card, and USB ports may remain powered, to receive input from the user. ACPI is a successor to an earlier Intel-Microsoft standard called Advanced Power Management, which allows a computer's BIOS to control power management functions.
Some programs allow the user to manually adjust the voltages supplied to the CPU, reducing the amount of electricity used by the CPU while it's on. Since many CPUs have "safety-nets" on either side of the spectrum (+/- the voltage parameters of a given CPU), one can reduce the voltage the processor uses (undervolting), reducing both the amount of heat produced and the amount of electricity consumed. Some CPUs automatically adjust the processor voltages and clock speed depending on the workload. This technology is called "SpeedStep" with intel processors, "PowerNow!"/"Cool'n'Quiet" with AMD chips, LongHaul with VIA CPUs, and LongRun with Transmeta processors.
In 2007, Intel Corporation released a Linux utility called PowerTOP, which measures and reports a PC's power consumption.
Performance Computers
A low power Alix.1C Mini-ITX embedded board with AMD Geode LX 800 together with Compact Flash, miniPCI and PCI slots, 44-pin IDE interface and 256MB RAM
A low power Alix.1C Mini-ITX embedded board with AMD Geode LX 800 together with Compact Flash, miniPCI and PCI slots, 44-pin IDE interface and 256MB RAM
As of 2007, several personal computer vendors (e.g., Everex, Linutop, Systemax, Zonbu and OLPC) ship dedicated low-power PCs. These systems provide minimal hardware peripherals and low performance processors, which makes them impractical for applications that require a lot of processing power such as computer gaming and video production. A low-power PCs is usually much smaller than traditional desktop. The limited capacity for upgrades, low performance and proprietary may lead to shorter lifespans and greater difficulty in repair.
Older laptops may provide similar performance with low power consumption. Reusing second-hand laptops may be an even more energy and material efficient alternative to such systems.
Routers, such as those compatible with the Linksys WRT54G, may be adapted for use in low power applications using replacement firmware.
Storage
Smaller form factor hard drives often consume less power than larger drives.
Unlike traditional hard disk drives, Solid-state drives store persistent data in either flash memory or DRAM. With no spindle motor or data platters to spin, power consumption may be reduced for low capacity devices.[13][14] Even at modest sizes, DRAM based SSDs may use more power than hard disks, e.g. i-RAM. Flash based solid state drives generally allow far fewer write cycles than hard drives. Storage capacity of solid state drives is also much more limited than for hard disks. Their shorter lifetime may make them less energy and material efficient in some applications.
Display
LCD monitors typically use a cold-cathode fluorescent bulb to provide light for the display. Some newer displays use an array of light-emitting diodes (LEDs) in place of the fluorescent bulb, which reduces the amount of electricity used by the display.[15]
Power supply
Desktop computer power supplies (PSUs) are generally 70–75% efficient[16], dissipating the remaining 30-25% of energy used as heat. An industry initiative called 80 PLUS certifies PSUs that are at least 80% efficient. Typically these models are drop-in replacements for older, less efficient PSUs of the same form factor. As of 2007-07-20, all new Energy Star 4.0-certified desktop PSUs must be at least 80% efficient.[17]
Communication
Network equipment manufacturers are developing network switches and routers that reduce operating costs as well.[18]
Materials recycling
Obsolete, but still functional computer systems can be repurposed, reused or donated [19] to various charities and non-profit organizations. Many charities have minimum system requirements for acceptable computer systems.[20] Parts may be salvaged from broken systems or those too old to be useful to charities, and the remains can be recycled through some retail outlets[21][22] and municipal or private recycling centers. Sometimes there is a charge for recycling, sometimes the cost is passed back to the manufacturers.
Recycling computing equipment keeps harmful materials such as lead, mercury, and hexavalent chromium out of landfills. However, oftentimes computers gathered through recycling drives are shipped to developing countries, where environmental standards are less strict than in North America and Europe[23]. The Silicon Valley Toxics Coalition estimates that 80% of the post-consumer e-waste collected for recycling is shipped abroad to countries such as China, India, and Pakistan.[24]
Computing supplies, such as printer cartridges, paper, and batteries may be recycled, and using recycled supplies helps close the loop.
Worker mobility
IT is increasingly able to influence the environmental impact of company operations outside the data center.
Voice over IP (VoIP) reduces the telephony wiring infrastructure by sharing the existing Ethernet copper (a toxic metal). VoIP and phone extension mobility also made hotelling of office space more practical.
The average annual energy consumption for U.S. office buildings is over 23 kilowatt hours per square foot; space heat, air conditioning and lighting together account for 70% of all energy consumed in a typical office building. [25] Hotelling reduces the square footage per employee because workers reserve space only when they need it. For many jobs -- sales, consulting, field service -- a dedicated office does not sit vacant, consuming energy for lighting and cooling.
Telecommuting can reduce space requirements, and the emissions caused by commuters. Telephony technologies have made it practical to operate whole departments outside the building: Call centers hire at-home agents whose physical absence from the building is practically indiscernible to customers.
Teleconferencing and telepresence technologies have the potential to cut down on business travel, a major source of greenhouse gas emissions.
Power generation
See also: Green energy
Computers can be powered by different sources, including windmills, hydroelectric, photovoltaic panels, or nuclear. Some of these sources are perceived to be more environmentally friendly than others. Other novel sources include human power, such as the Twibright Exciter project.
Virtualization
Main article: x86 virtualization
See also: Comparison of virtual machines
Computer virtualization is the process of running two or more logical computer systems on one set of physical hardware. The concept originated with the mainframe operating systems of the 1960s, but was commercialized for x86-compatible computers only in the 1990s. With virtualization, a system administrator could combine several physical systems into virtual machines on one single, powerful system, thereby unplugging the original hardware and reducing power and cooling consumption. Several commercial companies and open-source projects now offer software packages to enable a transition to virtual computing. Intel Corporation and AMD have also built proprietary virtualization enhancements to the x86 instruction set into each of their CPU product lines, in order to facilitate virtualized computing.
Power management
Main article: Power management
An open industry standard called Advanced Configuration and Power Interface (ACPI) provides a standard programming interface that allows an operating system to directly control the power saving aspects of the hardware. This allows the system to automatically turn off components such as monitors and hard drives after set periods of inactivity. In addition, a system may hibernate, in which it turns off nearly all components, including the CPU and the system RAM, greatly reducing the system's electricity usage. To resume from this state, some components, such as the keyboard, network interface card, and USB ports may remain powered, to receive input from the user. ACPI is a successor to an earlier Intel-Microsoft standard called Advanced Power Management, which allows a computer's BIOS to control power management functions.
Some programs allow the user to manually adjust the voltages supplied to the CPU, reducing the amount of electricity used by the CPU while it's on. Since many CPUs have "safety-nets" on either side of the spectrum (+/- the voltage parameters of a given CPU), one can reduce the voltage the processor uses (undervolting), reducing both the amount of heat produced and the amount of electricity consumed. Some CPUs automatically adjust the processor voltages and clock speed depending on the workload. This technology is called "SpeedStep" with intel processors, "PowerNow!"/"Cool'n'Quiet" with AMD chips, LongHaul with VIA CPUs, and LongRun with Transmeta processors.
In 2007, Intel Corporation released a Linux utility called PowerTOP, which measures and reports a PC's power consumption.
Performance Computers
A low power Alix.1C Mini-ITX embedded board with AMD Geode LX 800 together with Compact Flash, miniPCI and PCI slots, 44-pin IDE interface and 256MB RAM
A low power Alix.1C Mini-ITX embedded board with AMD Geode LX 800 together with Compact Flash, miniPCI and PCI slots, 44-pin IDE interface and 256MB RAM
As of 2007, several personal computer vendors (e.g., Everex, Linutop, Systemax, Zonbu and OLPC) ship dedicated low-power PCs. These systems provide minimal hardware peripherals and low performance processors, which makes them impractical for applications that require a lot of processing power such as computer gaming and video production. A low-power PCs is usually much smaller than traditional desktop. The limited capacity for upgrades, low performance and proprietary may lead to shorter lifespans and greater difficulty in repair.
Older laptops may provide similar performance with low power consumption. Reusing second-hand laptops may be an even more energy and material efficient alternative to such systems.
Routers, such as those compatible with the Linksys WRT54G, may be adapted for use in low power applications using replacement firmware.
Storage
Smaller form factor hard drives often consume less power than larger drives.
Unlike traditional hard disk drives, Solid-state drives store persistent data in either flash memory or DRAM. With no spindle motor or data platters to spin, power consumption may be reduced for low capacity devices.[13][14] Even at modest sizes, DRAM based SSDs may use more power than hard disks, e.g. i-RAM. Flash based solid state drives generally allow far fewer write cycles than hard drives. Storage capacity of solid state drives is also much more limited than for hard disks. Their shorter lifetime may make them less energy and material efficient in some applications.
Display
LCD monitors typically use a cold-cathode fluorescent bulb to provide light for the display. Some newer displays use an array of light-emitting diodes (LEDs) in place of the fluorescent bulb, which reduces the amount of electricity used by the display.[15]
Power supply
Desktop computer power supplies (PSUs) are generally 70–75% efficient[16], dissipating the remaining 30-25% of energy used as heat. An industry initiative called 80 PLUS certifies PSUs that are at least 80% efficient. Typically these models are drop-in replacements for older, less efficient PSUs of the same form factor. As of 2007-07-20, all new Energy Star 4.0-certified desktop PSUs must be at least 80% efficient.[17]
Communication
Network equipment manufacturers are developing network switches and routers that reduce operating costs as well.[18]
Materials recycling
Obsolete, but still functional computer systems can be repurposed, reused or donated [19] to various charities and non-profit organizations. Many charities have minimum system requirements for acceptable computer systems.[20] Parts may be salvaged from broken systems or those too old to be useful to charities, and the remains can be recycled through some retail outlets[21][22] and municipal or private recycling centers. Sometimes there is a charge for recycling, sometimes the cost is passed back to the manufacturers.
Recycling computing equipment keeps harmful materials such as lead, mercury, and hexavalent chromium out of landfills. However, oftentimes computers gathered through recycling drives are shipped to developing countries, where environmental standards are less strict than in North America and Europe[23]. The Silicon Valley Toxics Coalition estimates that 80% of the post-consumer e-waste collected for recycling is shipped abroad to countries such as China, India, and Pakistan.[24]
Computing supplies, such as printer cartridges, paper, and batteries may be recycled, and using recycled supplies helps close the loop.
Worker mobility
IT is increasingly able to influence the environmental impact of company operations outside the data center.
Voice over IP (VoIP) reduces the telephony wiring infrastructure by sharing the existing Ethernet copper (a toxic metal). VoIP and phone extension mobility also made hotelling of office space more practical.
The average annual energy consumption for U.S. office buildings is over 23 kilowatt hours per square foot; space heat, air conditioning and lighting together account for 70% of all energy consumed in a typical office building. [25] Hotelling reduces the square footage per employee because workers reserve space only when they need it. For many jobs -- sales, consulting, field service -- a dedicated office does not sit vacant, consuming energy for lighting and cooling.
Telecommuting can reduce space requirements, and the emissions caused by commuters. Telephony technologies have made it practical to operate whole departments outside the building: Call centers hire at-home agents whose physical absence from the building is practically indiscernible to customers.
Teleconferencing and telepresence technologies have the potential to cut down on business travel, a major source of greenhouse gas emissions.
Regulations of green Computing
Government
Many governmental agencies have implemented standards and regulations that encourage green computing. The U.S. Environmental Protection Agency's Energy Star program, which was launched in 1992, was revised in October 2006 to include stricter efficiency requirements for computer equipment, and a tiered ranking system for approved products.[2][3] In Europe, the Swedish Confederation of Professional Employees certifies personal computers, monitors, and other office equipment that meets the ergonomics, energy usage, emissions, and hazardous substances requirements of the TCO Certification program. The European Union's directives 2002/95/EC (RoHS), on the reduction of hazardous substances, and 2002/96/EC (WEEE) on waste electrical and electronic equipment required the substitution of heavy metals and flame retardants like PBBs and PBDEs in all electronic equipment put on the market starting on 2006-07-01. The directives placed responsibility on manufacturers for the gathering and recycling of old equipment.
In 2003, the California State Senate enacted the Electronic Waste Recycling Act, which establishes a state-wide recycling program for obsolete computer and consumer electronics equipment.[4] To pay for the program, the Act imposes a fee for each unit sold at retail, based on the size of the display.[5] The Act also establishes restrictions on hazardous substances, equal to the European Union's RoHS Directive.[6]
In 2008, a report published in the UK by the Department for Communities and Local Government, quantified that the potential carbon savings from increasing the usage of online public service delivery were significantly in excess of the negative impact of extra IT server capacity.[7]
Industry
* The Green Electronics Council offers the Electronic Products Environmental Assesment Tool (EPEAT) to assist in the purchase of "green" computing systems. The Council evaluates computing equipment on 28 criteria that measure a product's efficiency and sustainability attributes. On 2007-01-24, President George W. Bush issued Executive Order 13423, which requires all United States Federal agencies to use EPEAT when purchasing computer systems.[8][9]
* The Green Grid is a global consortium dedicated to advancing energy efficiency in data centers and business computing ecosystems. It was founded in February 2007 by several key companies in the industry – AMD, APC, Dell, HP, IBM, Intel, Microsoft, Rackable Systems, SprayCool, Sun Microsystems and VMware. The Green Grid has since grown to hundreds of members, including end users and government organizations, all focused on improving data center efficiency.
* IBM Project Big Green is a 1 billion-USD-per-year effort by IBM to design and promote energy efficiency in corporate data centers.[10] Launched in May 2007.
* Climate Savers Computing Initiative (CSCI) is an effort to reduce the electric power consumption of PCs in active and inactive states.[11] The CSCI provides a catalog of green products from it's member organizations, and information for reducing PC power consumption. It was started on 2007-06-12. The name stems from the World Wildlife Fund's Climate Savers program, which was launched in 1999.[12] The WWF is also a member of the Computing Initiative.
Many governmental agencies have implemented standards and regulations that encourage green computing. The U.S. Environmental Protection Agency's Energy Star program, which was launched in 1992, was revised in October 2006 to include stricter efficiency requirements for computer equipment, and a tiered ranking system for approved products.[2][3] In Europe, the Swedish Confederation of Professional Employees certifies personal computers, monitors, and other office equipment that meets the ergonomics, energy usage, emissions, and hazardous substances requirements of the TCO Certification program. The European Union's directives 2002/95/EC (RoHS), on the reduction of hazardous substances, and 2002/96/EC (WEEE) on waste electrical and electronic equipment required the substitution of heavy metals and flame retardants like PBBs and PBDEs in all electronic equipment put on the market starting on 2006-07-01. The directives placed responsibility on manufacturers for the gathering and recycling of old equipment.
In 2003, the California State Senate enacted the Electronic Waste Recycling Act, which establishes a state-wide recycling program for obsolete computer and consumer electronics equipment.[4] To pay for the program, the Act imposes a fee for each unit sold at retail, based on the size of the display.[5] The Act also establishes restrictions on hazardous substances, equal to the European Union's RoHS Directive.[6]
In 2008, a report published in the UK by the Department for Communities and Local Government, quantified that the potential carbon savings from increasing the usage of online public service delivery were significantly in excess of the negative impact of extra IT server capacity.[7]
Industry
* The Green Electronics Council offers the Electronic Products Environmental Assesment Tool (EPEAT) to assist in the purchase of "green" computing systems. The Council evaluates computing equipment on 28 criteria that measure a product's efficiency and sustainability attributes. On 2007-01-24, President George W. Bush issued Executive Order 13423, which requires all United States Federal agencies to use EPEAT when purchasing computer systems.[8][9]
* The Green Grid is a global consortium dedicated to advancing energy efficiency in data centers and business computing ecosystems. It was founded in February 2007 by several key companies in the industry – AMD, APC, Dell, HP, IBM, Intel, Microsoft, Rackable Systems, SprayCool, Sun Microsystems and VMware. The Green Grid has since grown to hundreds of members, including end users and government organizations, all focused on improving data center efficiency.
* IBM Project Big Green is a 1 billion-USD-per-year effort by IBM to design and promote energy efficiency in corporate data centers.[10] Launched in May 2007.
* Climate Savers Computing Initiative (CSCI) is an effort to reduce the electric power consumption of PCs in active and inactive states.[11] The CSCI provides a catalog of green products from it's member organizations, and information for reducing PC power consumption. It was started on 2007-06-12. The name stems from the World Wildlife Fund's Climate Savers program, which was launched in 1999.[12] The WWF is also a member of the Computing Initiative.
Origins and Rationale
Origins
In 1992, the U.S. Environmental Protection Agency launched Energy Star, a voluntary labeling program, designed to promote and recognize energy-efficiency in monitors, climate control equipment, and other technologies. This resulted in, among other things, the widespread adoption of sleep mode among consumer electronics. The term "green computing" was probably coined shortly after the Energy Star program began, and generally referred to power consumption-related issues. There are several USENET posts dating back to 1992 which use the term in this manner.[1] Also in 1992, the Swedish organization TCO Development launched the TCO Certification program to promote low magnetic and electrical emissions from CRT-based computer displays. The program was later expanded to include criteria on energy consumption, ergonomics, and the use of hazardous materials in construction.
Fueled by recent trends towards sustainability, the modern use of the term refers to systematic approaches to using computing technology efficiently. These include items such as addressing electronic waste, regulatory compliance, telecommuting policies, virtualization of server resources, cost accounting of energy use, thin client solutions, and many others.
[edit] Rationale
There are different philosophies related the implementation of green computing solutions. Virtually all proponents believe that businesses need to reevaluate some of their technological solutions and/or policies.
* Tactical incrementalists tend to preserve existing IT infrastructure and policies but will incorporate some environmental and social factors into future decisions. An example of this is to encourage power management policies for computing equipment. Changes of this type are generally easy to implement, are not part of a unified plan, and require little political effort. Similarly to a standard business model, fiscal viability remains the primary concern, with little or no value being attributed to the environmental or social factors.
* Strategic leaders recognize that environmental and social factors are contributing to a trend in disruptive technology. As such, there is considerable value to altering the existing infrastructure and policies. While the primary rationale is still cost efficiency, other factors such as branding, marketing, or hiring may be referenced as well. A baseline study will probably be conducted, and more complex, pronounced initiatives will be implemented. For example, the IT department may be held accountable for the cost of all electricity from computer related equipment, or desktop computing will be replaced with thin client computing.
* Deep green technologists believe that computing has reached a point of diminishing returns and needs to be substantially reevaluated. Unlike the other categories, social and environmental factors are construed as being on equal footing with fiscal responsibility. For example, a deep green technologist might implement a carbon offset policy to neutralize the cost of electricity for computing, or use solar powered web hosting services.
In 1992, the U.S. Environmental Protection Agency launched Energy Star, a voluntary labeling program, designed to promote and recognize energy-efficiency in monitors, climate control equipment, and other technologies. This resulted in, among other things, the widespread adoption of sleep mode among consumer electronics. The term "green computing" was probably coined shortly after the Energy Star program began, and generally referred to power consumption-related issues. There are several USENET posts dating back to 1992 which use the term in this manner.[1] Also in 1992, the Swedish organization TCO Development launched the TCO Certification program to promote low magnetic and electrical emissions from CRT-based computer displays. The program was later expanded to include criteria on energy consumption, ergonomics, and the use of hazardous materials in construction.
Fueled by recent trends towards sustainability, the modern use of the term refers to systematic approaches to using computing technology efficiently. These include items such as addressing electronic waste, regulatory compliance, telecommuting policies, virtualization of server resources, cost accounting of energy use, thin client solutions, and many others.
[edit] Rationale
There are different philosophies related the implementation of green computing solutions. Virtually all proponents believe that businesses need to reevaluate some of their technological solutions and/or policies.
* Tactical incrementalists tend to preserve existing IT infrastructure and policies but will incorporate some environmental and social factors into future decisions. An example of this is to encourage power management policies for computing equipment. Changes of this type are generally easy to implement, are not part of a unified plan, and require little political effort. Similarly to a standard business model, fiscal viability remains the primary concern, with little or no value being attributed to the environmental or social factors.
* Strategic leaders recognize that environmental and social factors are contributing to a trend in disruptive technology. As such, there is considerable value to altering the existing infrastructure and policies. While the primary rationale is still cost efficiency, other factors such as branding, marketing, or hiring may be referenced as well. A baseline study will probably be conducted, and more complex, pronounced initiatives will be implemented. For example, the IT department may be held accountable for the cost of all electricity from computer related equipment, or desktop computing will be replaced with thin client computing.
* Deep green technologists believe that computing has reached a point of diminishing returns and needs to be substantially reevaluated. Unlike the other categories, social and environmental factors are construed as being on equal footing with fiscal responsibility. For example, a deep green technologist might implement a carbon offset policy to neutralize the cost of electricity for computing, or use solar powered web hosting services.
Green computing
Green computing is the study and practice of using computing resources efficiently. Typically, green computing systems or products take into account the so-called triple bottom line of people, planet, profit. This differs somewhat from traditional or standard business practices that focus mainly on the economic viability of a computing solution. These focuses are similar to those of green chemistry; reduction of the use of hazardous materials such as lead at the manufacturing and recycling stages, maximized energy efficiency during the product's lifetime, and recyclability or biodegradability of both a defunct product and of any factory waste.
A typical green computing solution attempts to address some or all of these factors by using environmentally friendly products in an efficient system. For example, an IT manager might purchase Electronic Products Environmental Assessment Tool (EPEAT)-approved hardware combined with a thin client solution. As compared to a traditional desktop PC configuration, such a configuration would probably reduce IT maintenance-related activities, extend the useful life of the hardware, and allow for responsible recycling of the equipment past its useful life.
The market for Green IT products is growing fastest at the enterprise computing level.[citation needed] Power and cooling costs are some of the biggest concerns addressed by green computing solutions
A typical green computing solution attempts to address some or all of these factors by using environmentally friendly products in an efficient system. For example, an IT manager might purchase Electronic Products Environmental Assessment Tool (EPEAT)-approved hardware combined with a thin client solution. As compared to a traditional desktop PC configuration, such a configuration would probably reduce IT maintenance-related activities, extend the useful life of the hardware, and allow for responsible recycling of the equipment past its useful life.
The market for Green IT products is growing fastest at the enterprise computing level.[citation needed] Power and cooling costs are some of the biggest concerns addressed by green computing solutions
Autonomic Computing
The approaches to implement intelligent systems can be classified into those of biological organisms, silicon automata, and computing systems. Based on CI studies, autonomic computing [Wang, 2004] is proposed as a new and advanced computing technique built upon the routine, algorithmic, and adaptive systems. The approaches to computing can be classified into two categories known as imperative and autonomic computing. Corresponding to these, computing systems may be implemented as imperative or autonomic computing systems.
Definition 8. An imperative computing system is a passive system that implements deterministic, context-free, and stored-program controlled behaviors.
Definition 9. An autonomic computing system is an intelligent system that autonomously carries out robotic and interactive actions based on goal- and event-driven mechanisms.
The imperative computing system is a traditional passive system that implements deterministic, context-free, and stored-program controlled behaviors, where a behavior is defined as a set of observable actions of a given computing system. The autonomic computing system is an active system that implements nondeterministic, context-dependent, and adaptive behaviors, which do not rely on instructive and procedural information, but are dependent on internal status and willingness that formed by long-term historical events and current rational or emotional goals.
The first three categories of computing techniques as shown in Table 3 are imperative. In contrast, the autonomic computing systems are an active system that implements nondeterministic, context-sensitive, and adaptive behaviors. Autonomic computing does not rely on imperative and procedural instructions, but are dependent on perceptions and inferences based on internal goals as revealed in CI.
Definition 8. An imperative computing system is a passive system that implements deterministic, context-free, and stored-program controlled behaviors.
Definition 9. An autonomic computing system is an intelligent system that autonomously carries out robotic and interactive actions based on goal- and event-driven mechanisms.
The imperative computing system is a traditional passive system that implements deterministic, context-free, and stored-program controlled behaviors, where a behavior is defined as a set of observable actions of a given computing system. The autonomic computing system is an active system that implements nondeterministic, context-dependent, and adaptive behaviors, which do not rely on instructive and procedural information, but are dependent on internal status and willingness that formed by long-term historical events and current rational or emotional goals.
The first three categories of computing techniques as shown in Table 3 are imperative. In contrast, the autonomic computing systems are an active system that implements nondeterministic, context-sensitive, and adaptive behaviors. Autonomic computing does not rely on imperative and procedural instructions, but are dependent on perceptions and inferences based on internal goals as revealed in CI.
Applications of CI
Applications of CI
A wide range of applications of CI has been identified in multidisciplinary and transdisciplinary areas, such as: (1) The architecture of future generation computers; (2) Estimation the capacity of human memory; (3) Autonomic computing; (4) Simulation of human cognitive behaviors using denotational mathematics.
The Architecture of Future Generation Computers
Conventional machines are invented to extend human physical capability, while modern information processing machines, such as computers, communication networks, and robots, are developed for extending human intelligence, memory, and the capacity for information processing [Wang, 2004]. Recent advances in CI provide formal description of an entire set of cognitive processes of the brain [Wang et al., 2006]. The fundamental research in CI also creates an enriched set of contemporary denotational mathematics [Wang, 2006c], for dealing with the extremely complicated objects and problems in natural intelligence, neural informatics, and knowledge manipulation.
The theory and philosophy behind the next generation computers and computing methodologies are CI [Wang, 2004]. It is commonly believed that the future-generation computers, known as the cognitive computers, will adopt non-von Neumann [von Neumann, 1946] architectures. The key requirements for implementing a conventional stored-program controlled computer are the generalization of common computing architectures and the computer is able to interpret the data loaded in memory as computing instructions. These are the essences of stored-program controlled computers known as the von Neumann architecture, which encompasses five essential components to implement general-purpose programmable digital computers.
Definition 4. A von Neumann Architecture (VNA) of computers is a 5-tuple that consists of the components: (a) the arithmetic-logic unit (ALU), (b) the control unit (CU) with a program counter (PC), (c) a memory (M), (d) a set of input/output (I/O) devices, and (e) a bus (B) that provides the data path between these components.
Definition 5. Conventional computers with VNA are aimed at stored-program-controlled data processing based on mathematical logic and Boolean algebra.
A VNA computer is centric by the bus and characterized by the all purpose memory for both data and instructions. A VNA machine is an extended Turing machine (TM), where the power and functionality of all components of TM including the control unit (with wired instructions), the tape (memory), and the head of I/O, are greatly enhanced and extended with more powerful instructions and I/O capacity.
Definition 6. A Wang Architecture (WA) of computers, known as the Cognitive Computers is a parallel structure encompassing an Inference Engine (IE) and a Perception Engine (PE) [Wang, 2006a].
The future generation cognitive computers based on WA are not centered by a CPU for data manipulation as the VNA computers do. The WA computers are centered by the concurrent IE and PE for cognitive learning and autonomic perception based on abstract concept inferences and empirical stimuli perception. The IE is designed for concept/knowledge manipulation according to concept algebra [Wang, 2006d], particularly the 9 concept operations for knowledge acquisition, creation, and manipulation. The PE is designed for sensory and perception processing according to RTPA [Wang, 2002b] and the formally described cognitive process models of the perception layers as defined in the LRMB model [Wang et al., 2006].
Definition 7. Cognitive computers with WA are aimed at cognitive and perceptive concept/knowledge processing based on contemporary denotational mathematics, i.e. concept algebra, Real-Time Process Algebra (RTPA), and system algebra.
As that of mathematical logic and Boolean algebra are the mathematical foundations of VNA computers. The mathematical foundations of WA computers are based on denotational mathematics [Wang, 2006b, 2006c]. As described in the LRMB reference model [Wang et al., 2006], since all the 39 fundamental cognitive processes of human brains can be formally described in CA and RTPA [Wang, 2002b, 2006e]. In other words, they are simulatable and executable by the WA-based cognitive computers.
Estimation the Capacity of Human Memory
Despite the fact that the number of neurons in the brain has been identified in cognitive and neural sciences, the magnitude of human memory capacity is still unknown. According to the Object-Attribute-Relation (OAR) model [Wang, 2007c], a recent discovery in CI is that the upper bound of memory capacity of the human brain is in the order of 108,432 bits [Wang et al., 2003]. The determination of the magnitude of human memory capacity is not only theoretically significant in CI, but also practically useful to explain the human potential, as well as the gaps between the natural and machine intelligence. This work indicates that the next generation computer memory systems may be built according to the OAR model rather than the traditional container metaphor, because the former is more powerful, flexible, and efficient to generate a tremendous memory capacity by using limited number of neurons in the brain or hardware cells in the next generation computers.
A wide range of applications of CI has been identified in multidisciplinary and transdisciplinary areas, such as: (1) The architecture of future generation computers; (2) Estimation the capacity of human memory; (3) Autonomic computing; (4) Simulation of human cognitive behaviors using denotational mathematics.
The Architecture of Future Generation Computers
Conventional machines are invented to extend human physical capability, while modern information processing machines, such as computers, communication networks, and robots, are developed for extending human intelligence, memory, and the capacity for information processing [Wang, 2004]. Recent advances in CI provide formal description of an entire set of cognitive processes of the brain [Wang et al., 2006]. The fundamental research in CI also creates an enriched set of contemporary denotational mathematics [Wang, 2006c], for dealing with the extremely complicated objects and problems in natural intelligence, neural informatics, and knowledge manipulation.
The theory and philosophy behind the next generation computers and computing methodologies are CI [Wang, 2004]. It is commonly believed that the future-generation computers, known as the cognitive computers, will adopt non-von Neumann [von Neumann, 1946] architectures. The key requirements for implementing a conventional stored-program controlled computer are the generalization of common computing architectures and the computer is able to interpret the data loaded in memory as computing instructions. These are the essences of stored-program controlled computers known as the von Neumann architecture, which encompasses five essential components to implement general-purpose programmable digital computers.
Definition 4. A von Neumann Architecture (VNA) of computers is a 5-tuple that consists of the components: (a) the arithmetic-logic unit (ALU), (b) the control unit (CU) with a program counter (PC), (c) a memory (M), (d) a set of input/output (I/O) devices, and (e) a bus (B) that provides the data path between these components.
Definition 5. Conventional computers with VNA are aimed at stored-program-controlled data processing based on mathematical logic and Boolean algebra.
A VNA computer is centric by the bus and characterized by the all purpose memory for both data and instructions. A VNA machine is an extended Turing machine (TM), where the power and functionality of all components of TM including the control unit (with wired instructions), the tape (memory), and the head of I/O, are greatly enhanced and extended with more powerful instructions and I/O capacity.
Definition 6. A Wang Architecture (WA) of computers, known as the Cognitive Computers is a parallel structure encompassing an Inference Engine (IE) and a Perception Engine (PE) [Wang, 2006a].
The future generation cognitive computers based on WA are not centered by a CPU for data manipulation as the VNA computers do. The WA computers are centered by the concurrent IE and PE for cognitive learning and autonomic perception based on abstract concept inferences and empirical stimuli perception. The IE is designed for concept/knowledge manipulation according to concept algebra [Wang, 2006d], particularly the 9 concept operations for knowledge acquisition, creation, and manipulation. The PE is designed for sensory and perception processing according to RTPA [Wang, 2002b] and the formally described cognitive process models of the perception layers as defined in the LRMB model [Wang et al., 2006].
Definition 7. Cognitive computers with WA are aimed at cognitive and perceptive concept/knowledge processing based on contemporary denotational mathematics, i.e. concept algebra, Real-Time Process Algebra (RTPA), and system algebra.
As that of mathematical logic and Boolean algebra are the mathematical foundations of VNA computers. The mathematical foundations of WA computers are based on denotational mathematics [Wang, 2006b, 2006c]. As described in the LRMB reference model [Wang et al., 2006], since all the 39 fundamental cognitive processes of human brains can be formally described in CA and RTPA [Wang, 2002b, 2006e]. In other words, they are simulatable and executable by the WA-based cognitive computers.
Estimation the Capacity of Human Memory
Despite the fact that the number of neurons in the brain has been identified in cognitive and neural sciences, the magnitude of human memory capacity is still unknown. According to the Object-Attribute-Relation (OAR) model [Wang, 2007c], a recent discovery in CI is that the upper bound of memory capacity of the human brain is in the order of 108,432 bits [Wang et al., 2003]. The determination of the magnitude of human memory capacity is not only theoretically significant in CI, but also practically useful to explain the human potential, as well as the gaps between the natural and machine intelligence. This work indicates that the next generation computer memory systems may be built according to the OAR model rather than the traditional container metaphor, because the former is more powerful, flexible, and efficient to generate a tremendous memory capacity by using limited number of neurons in the brain or hardware cells in the next generation computers.
Subscribe to:
Posts (Atom)
