Showing posts with label energy usage. Show all posts
Showing posts with label energy usage. Show all posts
Saturday, April 30, 2016
Blades and the Unified Fabric
Converging your enterprise network(s) onto a single platform may seem like a no-brainer. After all, who wouldn't want to manage and maintain a single network rather than separate data, storage and even voice architectures?
But beyond the major question of which format to build on, there are a number of other aspects to take into consideration, such as what sort of server architecture will best fit into this unified environment.
Organizations backing blade servers have made no secret of their intention to promote their technology as the ideal solution. Besides their ability to aggregate network ports, the newest blade architectures will be equipped to handle all of the single-wire formats under consideration: 10 GbE, FCoE, iSCSI and Infiniband.
The Blade.org group, consisting of BLADE Network Technologies, Brocade, Emulex, IBM, Mellanox and other vendors, just released a white paper entitled "Blade Platforms and Network Convergence" highlighting all the benefits of convergence around the blade, including the scrapping of multiple adapters and switch ports and the ability to institute dynamic allocation of end-to-end resources.
Anything is better than the hodgepodge of hardware and software that currently makes up the data center, according to IDC's Richard Villars. He tells ServerWatch that the continued practice of adding servers wherever they happen to fit in the racks and then linking it all to storage somewhere else is unsustainable. There may be differences in the way vendors are building their unified platforms, but they should all provide significant improvements in network simplicity.
Cisco may be going one better by bringing in its own blade design. The company is rumored to be working on an Intel-based device code-named the "California Server" with plans to link it closely to the Nexus switch. The company has not commented on the story yet, but it would seem to make a lot of sense for Cisco to use its considerable resources to devise a blade to go with its network technology, even if it does cross swords with some of the company's major partners like HP and IBM.
Blades, of course, have the advantage of providing high-density environments in an industry that is usually cramped for space to begin with. But they by no means have a lock on that feature. Processor points out in a recent editorial that some of the new 1-inch chassis are starting to push back by incorporating multiple quad-core CPUs, RAM sockets and drive bays.
More than likely, though, blades and standard servers will play leapfrog for a considerable period as the unified network concept spreads out across the industry. But no matter which platform is adopted (and indeed, why not both?) the goal is to provide a flexible, virtual environment for users - two primary ingredients for the cloud.
Read Art's article, "SAN Convergence We Can Live With" - http://www.itbusinessedge.com/blogs/atc/?p=541&nr=inbound
Article Source: http://EzineArticles.com/expert/Arthur_Cole/213479
Article Source: http://EzineArticles.com/1835917
Refurbished Servers HP: Selected Hardware Should Support Environmental and Usage Needs
Refurbished devices allow a business to purchase top rated hardware at a budget worthy price. Servers have become essential to businesses today due to both the data being handled and the various advanced technologies now available. The longstanding reputation of HP products has made them one of the top providers for high-end computing equipment. Smaller businesses have a hard time determining if a server is needed for their operations. This hardware is often confused with regular personal computers due to the hardware similarities. A server does have many of the same components as a typical computer but significant differences as well. Both contain memory, power supplies, cases, and central processing units. Advanced hardware components put these devices above PC's and make them a little harder to learn. PC's in most cases do not contain redundant components or hot swappable capabilities. Refurbished servers by HP allow a business to utilize advanced technology features without the big price tag of new equipment.
Used Servers: What Differences Exist Between Tower, Rack, and Blade Models?
Used servers come in three basic styles sometimes referred to as form factors. The form factor revolves around the standards and dimensions used to hold the computing hardware. Tower, blade, and rack designs may be chosen based on the particular needs of a business. Traditionally, tower models were the first type used among companies. It is this design that made the two confusing to both owners and those IT professionals working with them. Towers look much like a typical personal computer because they reside in an upright case and are a standalone machine. They work best in smaller data centers but their immense spacing requirements often prevent them from being used in larger network environments. These designs are also a little noisier than other forms. Cabling can be a real mess as well since it is spread out much like any other type of computer. A smaller business with five to twenty-five employees can benefit the most from tower designed used servers.
Racks hold a multitude of standardized servers which may be stacked vertically to save space. Additional uniform components can be acquired such as mountable switches to make this setup highly convenient. A 1U model is the standard meaning it only takes up one rack. 2U, 4U, and even 1/2U designs are also available for purchase. Smaller models work well for the typical business workload; however, larger designs can accommodate larger storage capabilities. Blade components cannot be interchanged between varying vendors as other designs. An HP blade server will experience large problems if combined with a Dell version or vice versa. Rack designs have the property of being self-contained whereas blade models do not have all the parts required to function in this same manner. Blade chassis designs vary among vendors and can include power and cooling units in addition to refurbished servers by HP. They hold numerous servers along with multiple support components not supplied within this data storage method. Every type has their advantages, making it essential for a business to be aware of what is available before making a buying decision.
Refurbished servers from HP may be purchased in vertically stacked or standalone designs. Used servers offer the same benefits but allow companies to utilize great hardware at a lower cost.
Article Source: http://EzineArticles.com/expert/Christine_Spiller/118197
Used Servers: What Differences Exist Between Tower, Rack, and Blade Models?
Used servers come in three basic styles sometimes referred to as form factors. The form factor revolves around the standards and dimensions used to hold the computing hardware. Tower, blade, and rack designs may be chosen based on the particular needs of a business. Traditionally, tower models were the first type used among companies. It is this design that made the two confusing to both owners and those IT professionals working with them. Towers look much like a typical personal computer because they reside in an upright case and are a standalone machine. They work best in smaller data centers but their immense spacing requirements often prevent them from being used in larger network environments. These designs are also a little noisier than other forms. Cabling can be a real mess as well since it is spread out much like any other type of computer. A smaller business with five to twenty-five employees can benefit the most from tower designed used servers.
Racks hold a multitude of standardized servers which may be stacked vertically to save space. Additional uniform components can be acquired such as mountable switches to make this setup highly convenient. A 1U model is the standard meaning it only takes up one rack. 2U, 4U, and even 1/2U designs are also available for purchase. Smaller models work well for the typical business workload; however, larger designs can accommodate larger storage capabilities. Blade components cannot be interchanged between varying vendors as other designs. An HP blade server will experience large problems if combined with a Dell version or vice versa. Rack designs have the property of being self-contained whereas blade models do not have all the parts required to function in this same manner. Blade chassis designs vary among vendors and can include power and cooling units in addition to refurbished servers by HP. They hold numerous servers along with multiple support components not supplied within this data storage method. Every type has their advantages, making it essential for a business to be aware of what is available before making a buying decision.
Refurbished servers from HP may be purchased in vertically stacked or standalone designs. Used servers offer the same benefits but allow companies to utilize great hardware at a lower cost.
Article Source: http://EzineArticles.com/expert/Christine_Spiller/118197
Friday, April 29, 2016
Monitoring Temperature And Humidity Inside The Server Rack
The increased utilization of blade servers has caused the control of heat and humidity in the server racks to become much more critical than previously. These units increase the density of the equipment in each rack and therefore the heat generated. Alarms are necessary at the rack level because a single server overheating will not raise the ambient room temperature enough to set off the room alarm sensors. These rack alarms can provide early problem warning and prevent server damage and the associated network problems in a timely fashion.
These alarms can be at the rack, or by email, or by SNMP traps, or by all three methods depending on the system. There are several different types of sensor systems from several different manufacturers, available at present. One of them is a compact environmental monitor with one onboard temperature sensor and in addition the ability to support up to 16 remote sensors. The system is very small so you don't have to change your system. If you receive a problem alert, you are able to identify the rack and power strip having the trouble, and turn off the individual outlet or reboot it.
Another power management and control system available is able to track temperature, humidity, airflow, and water at various locations from 15 to 100 feet away. This system needs only limited rack space and can control individual power receptacles. Another system on the market utilizes two sensor probes in each rack, one in front, one in back, to better measure the environment in the rack. This gives you a thermographic view of the temperature range in the rack and can be displayed in color for all the different racks in the computer center. This information can be used to determine the size and location of any additional A.C. or fans or blowers needed to correct a cooling system. It can also assist in the rearrangement, if necessary, of the rack layout and separation of a cold aisle/hot aisle system or any rack system, or when additional racks are added to a functioning system.
The control of heat generation and temperature in the data center continues to be one of the most critical areas. The monitoring of the environment has gone from the room, to the zone, to inside the individual server rack to give a much more precise picture of what is happening and allow better response to emergency conditions.
Server Racks and Cable Inc. http://www.serverracksandcable.com provides server rack solutions to Fortune 500 companies, governments, and military agencies.
Article Source: http://EzineArticles.com/expert/Mike_Jarrett/110141
Article Source: http://EzineArticles.com/684543
These alarms can be at the rack, or by email, or by SNMP traps, or by all three methods depending on the system. There are several different types of sensor systems from several different manufacturers, available at present. One of them is a compact environmental monitor with one onboard temperature sensor and in addition the ability to support up to 16 remote sensors. The system is very small so you don't have to change your system. If you receive a problem alert, you are able to identify the rack and power strip having the trouble, and turn off the individual outlet or reboot it.
Another power management and control system available is able to track temperature, humidity, airflow, and water at various locations from 15 to 100 feet away. This system needs only limited rack space and can control individual power receptacles. Another system on the market utilizes two sensor probes in each rack, one in front, one in back, to better measure the environment in the rack. This gives you a thermographic view of the temperature range in the rack and can be displayed in color for all the different racks in the computer center. This information can be used to determine the size and location of any additional A.C. or fans or blowers needed to correct a cooling system. It can also assist in the rearrangement, if necessary, of the rack layout and separation of a cold aisle/hot aisle system or any rack system, or when additional racks are added to a functioning system.
The control of heat generation and temperature in the data center continues to be one of the most critical areas. The monitoring of the environment has gone from the room, to the zone, to inside the individual server rack to give a much more precise picture of what is happening and allow better response to emergency conditions.
Server Racks and Cable Inc. http://www.serverracksandcable.com provides server rack solutions to Fortune 500 companies, governments, and military agencies.
Article Source: http://EzineArticles.com/expert/Mike_Jarrett/110141
Article Source: http://EzineArticles.com/684543
Wednesday, March 16, 2011
Efficient Energy Usage and Lower TCO for Blade Servers
With the launch of dual-core Xeon chip for blade servers, Intel has become the ability of businesses to cash in the form of low power consumption and greater reliability scored shop.
Studies indicate a server blade power and cooling costs 15-25 percent to meet the needs of our clients in question. Dual-core/dual-processor capabilities are ideal for a wide range of low-power communications and embedded applications.
blade server is: Even movies that are not less powerful processors, the total power and cooling requirements of a rack, eliminating many of the infrastructure components to reduce the energy consumed twice contained 1U server 2U rack.
use a blade server, not half or two thirds of the power of traditional 1U and 2U rack mount is the same computing power. The reason is that the consolidation of resources allows the blade to achieve economies of scale. For example, instead of 10 offers 500 watts for 10 traditional 1U servers, server blades can have two 2000 watts of power.
With all that support equipment is removed and the common chassis, the blade is much smaller. With the reduction in the blade server can be hot-plug power and cooling in the chassis. Therefore, the chassis offers more efficient space and power.
Add new processors with low energy efficiency of a single chassis and blade servers are an attractive option for server consolidation projects.
The three new Xeon server chips this year, published by Intel Sossaman (launched March 14), Dempsey (May 23) and Woodcrest (June 26).
The Xeon chips offer the advantages of combining the characteristics of two separate cores with intelligent power management for high yield performance much more per watt than previous platforms, single-core Intel Xeon-based. What follows is a brief explanation of each:
(Also known as low-voltage Xeon) Sossaman consumes about 30 watts and 31 watts with a thermal design power (TDP) of assessment. It integrates two to four times higher performance per watt than its predecessor, a version of a 55-watt single-core Intel Xeon chip called "Irwindale".
The processor is available in 1.66 and 2.0 GHz and offers speeds of 32-bit (since it is derived from a handheld device), instead of 64-bit addressing. It has a speed of 667 MHz frontside bus and L2 cache size of 2 MB.
In almost a month apart, were Dempsey (Dual-Core Xeon 5000 series) and Woodcrest (Xeon 5100) published by Intel. The launch of Woodcrest Dempsey makes some believe, based on Intel "Bensley" server was immediately necessary. However, Intel Dempsey is sold as a low-cost alternative to Woodcrest.
The comparison between the most important, with Dempsey Woodcrest Woodcrest consumes less power and less heat for Dempsey takes the micro-architecture has been optimized at the top is to reduce energy consumption. Woodcrest is available in three varieties, 40W TDP optimized versions for blades, the public 65W TDP parts, and are above 3 GHz and a TDP of 80W.
All the pieces fall below 3 GHz in the range of 65W or 40W TDP. In comparison, the top bin shares Dempsey has a TDP of 130W, and parts of the middle range were measured at 95 W, and the need for raw power and thermal problems have not prevented a knife Dempsey.
He also read Woodcrest and clock gating, the average improvement in performance to help reduce.
rate ranges from 2.67 to 3.73 GHz Dempsey Some models have a 667 MHz FSB, while others have a 1066 MHz FSB
Woodcrest, the first Intel Core microarchitecture to set up, 80 percent and increase performance while reducing energy consumption by 20 percent compared to the Pentium D.
It has a FSB of 1333 MHz, in most models, except 5110 and 5120, with a FSB of 1066 MHz, with the fastest processor at 3.0 GHz log Any use 5148LV Woodcrest LGA 771 and all have at least 5160 and a TDP of 65 W, which is much less than the previous generation of 130 W. 5160 has a TDP of 80 W. Watts, far less than 130
Intel has begun shipping to customers the next generation of Intel server, the server board and server chassis with dual-core Xeon 5000 and 5100 (Dempsey and Woodcrest).
The use of blade servers is advantageous for businesses large and small. For small companies, one or more individual servers often cost a small network should be supported. On large installations, in particular the expected increase capacity quickly, the blade server architecture offers increased efficiency and ease of scalability.
In summary, blade servers - using the latest versions of Intel Xeon dual-core - to increase reliability, reduce energy consumption, and are the logical choice for a company to save money and administration.
Studies indicate a server blade power and cooling costs 15-25 percent to meet the needs of our clients in question. Dual-core/dual-processor capabilities are ideal for a wide range of low-power communications and embedded applications.
blade server is: Even movies that are not less powerful processors, the total power and cooling requirements of a rack, eliminating many of the infrastructure components to reduce the energy consumed twice contained 1U server 2U rack.
use a blade server, not half or two thirds of the power of traditional 1U and 2U rack mount is the same computing power. The reason is that the consolidation of resources allows the blade to achieve economies of scale. For example, instead of 10 offers 500 watts for 10 traditional 1U servers, server blades can have two 2000 watts of power.
With all that support equipment is removed and the common chassis, the blade is much smaller. With the reduction in the blade server can be hot-plug power and cooling in the chassis. Therefore, the chassis offers more efficient space and power.
Add new processors with low energy efficiency of a single chassis and blade servers are an attractive option for server consolidation projects.
The three new Xeon server chips this year, published by Intel Sossaman (launched March 14), Dempsey (May 23) and Woodcrest (June 26).
The Xeon chips offer the advantages of combining the characteristics of two separate cores with intelligent power management for high yield performance much more per watt than previous platforms, single-core Intel Xeon-based. What follows is a brief explanation of each:
(Also known as low-voltage Xeon) Sossaman consumes about 30 watts and 31 watts with a thermal design power (TDP) of assessment. It integrates two to four times higher performance per watt than its predecessor, a version of a 55-watt single-core Intel Xeon chip called "Irwindale".
The processor is available in 1.66 and 2.0 GHz and offers speeds of 32-bit (since it is derived from a handheld device), instead of 64-bit addressing. It has a speed of 667 MHz frontside bus and L2 cache size of 2 MB.
In almost a month apart, were Dempsey (Dual-Core Xeon 5000 series) and Woodcrest (Xeon 5100) published by Intel. The launch of Woodcrest Dempsey makes some believe, based on Intel "Bensley" server was immediately necessary. However, Intel Dempsey is sold as a low-cost alternative to Woodcrest.
The comparison between the most important, with Dempsey Woodcrest Woodcrest consumes less power and less heat for Dempsey takes the micro-architecture has been optimized at the top is to reduce energy consumption. Woodcrest is available in three varieties, 40W TDP optimized versions for blades, the public 65W TDP parts, and are above 3 GHz and a TDP of 80W.
All the pieces fall below 3 GHz in the range of 65W or 40W TDP. In comparison, the top bin shares Dempsey has a TDP of 130W, and parts of the middle range were measured at 95 W, and the need for raw power and thermal problems have not prevented a knife Dempsey.
He also read Woodcrest and clock gating, the average improvement in performance to help reduce.
rate ranges from 2.67 to 3.73 GHz Dempsey Some models have a 667 MHz FSB, while others have a 1066 MHz FSB
Woodcrest, the first Intel Core microarchitecture to set up, 80 percent and increase performance while reducing energy consumption by 20 percent compared to the Pentium D.
It has a FSB of 1333 MHz, in most models, except 5110 and 5120, with a FSB of 1066 MHz, with the fastest processor at 3.0 GHz log Any use 5148LV Woodcrest LGA 771 and all have at least 5160 and a TDP of 65 W, which is much less than the previous generation of 130 W. 5160 has a TDP of 80 W. Watts, far less than 130
Intel has begun shipping to customers the next generation of Intel server, the server board and server chassis with dual-core Xeon 5000 and 5100 (Dempsey and Woodcrest).
The use of blade servers is advantageous for businesses large and small. For small companies, one or more individual servers often cost a small network should be supported. On large installations, in particular the expected increase capacity quickly, the blade server architecture offers increased efficiency and ease of scalability.
In summary, blade servers - using the latest versions of Intel Xeon dual-core - to increase reliability, reduce energy consumption, and are the logical choice for a company to save money and administration.
Subscribe to:
Posts (Atom)