• <More on Intel.com

Micro Servers Based On Intel® Xeon® Processors

Micro server. Macro performance per watt.

Achieve New Levels of Efficiency and Density for Emerging Scale-out Workloads

The micro server represents a new server architecture characterized by many lightweight server nodes bundled together in a shared chassis infrastructure. They’re designed specifically for:

  • Density
  • Lower power per node
  • Reduced costs
  • Increased operational efficiency

Micro servers are particularly well suited for lightweight scale-out of non-enterprise applications such as:

  • Simple Front End Web Tier
  • Low End Dedicated Hosting
  • Basic Content Delivery Node

Intel® Xeon® processor E3-1200 v2 product family-based micro servers offer improved performance per watt—up to 39% more energy efficiency versus the previous generation at the node level.1,2 For some less demanding scale-out applications, microarchitecture based on the Intel® Atom™ processor, code name Centerton (launching in the second half of 2012), could be an option. Data center-class capabilities, such as x86 compatibility, 64-bit architecture, ECC memory, virtualization support, integrated security, and scalability, will help you maximize node density, improve I/O performance, and minimize frustrations—all reasons why a micro server based on the Intel® Xeon® processor E3-1200 v2 product family is a simple and affordable solution for your data center.

Server System Infrastructure (SSI) Forum is a leading server industry group that drives server form factor standards including the Micro Module Server Specification. The Micro Module Server Specification provides a server board form factor and system interfaces specifically designed for the micro server segment. This standard will enable innovative system architectures to reduce product costs and increase product and data center efficiencies.

Micro Servers Provide Up To 3.6x Greater Performance Per Rack

A rack populated with micro servers based on the Intel® Xeon® processor E3-1200 v2 product family delivers up to 3.6x greater performance compared to a rack populated with 1U servers based on the Intel® Xeon® processor E3-1200 product family.3

Baseline, Intel® Xeon® processor E3-1200 Product Family

Baseline configuration: Intel® Xeon® processor E3-1220L 

 

  • Maximum number of nodes, 42U rack: 41 1U server + 1x 1U 48 port GbE Ethernet switch
  • Performance per node/rack: Best published SPECint*_rate_base2006 score of 69 as of April 26, 2012. http://www.spec.org/cpu2006/results/res2012q1/cpu2006-20111219-19191.html. 41 servers = 2829
  • Power per node: Max power consumption of one Intel® Xeon® processor E3-1220L on a Intel® C206 CRB using SPECpower_ssj2008, Enhanced Intel SpeedStep® Technology enabled, Intel® Turbo Boost Technology enabled, 8 GB memory (2x 4 GB DDR3-1333 UDIMM), 64G 3 Gb/s SATA SSD , Microsoft Windows* 2008 R2 SP1. Java SE Runtime Environment* (build 1.6.0_30-b12), Java HotSpot* 64-Bit Server VM (build 20.5-b03, mixed mode). Source: Intel internal testing as of Mar 2012. Score: (ssj_ops@100%: 147,345,  Power@100%: 50.1W, Active idle power: 25.4)
  • Power per rack: 2.3kW total. 1U switch = 240W, 41 1U server nodes = 2054W
Micro Server, Intel® Xeon® processor E3-1200 v2 Product Family

Micro server configuration: Intel® Xeon® processor E3-1220L v2

 

  • Maximum number of nodes: SSI rack = 10 3U chassis with 119 nodes + 3x 1U 48 port GbE switches
  • Performance per node/rack: Intel® C206 Qual Platform with one Intel® Xeon® processor E3-1220L v2 (3M cache, 2.30 GHz, L1 stepping), Intel SpeedStep® Technology enabled, Intel® Turbo Boost Technology enabled, Intel® Hyper-Threading Technology enabled, 16 GB memory (2x 8 GB DDR3-1600 ECC UDIMM), 160 GB SATA 7200RPM HDD, Red Hat* Enterprise Linux Server 6.2 with kernel: 2.6.32-220.el6.x86_64. Compiler version: 12.1.0.225 of Intel® C++ Studio XE and Intel® Fortran. Source: Intel internal testing as of Feb 2012. Score: SPECint_rate_base2006 of 85.7, 119 nodes = 10198.3
  • Power per node: Based on Intel estimates of 36W per node with one Intel® Xeon® processor E3-1220L v2 node, Intel SpeedStep® Technology enabled, Intel® Turbo Boost Technology enabled, 8 GB memory (2x 4 GB DDR3-1600 UDIMM), 1x SSD, assuming shared cooling resources and shared power supplies
  • Power per rack: 5.0kW total. 3x 1U switches at 240W per switch = 720W, 119 nodes at 36W estimated per node = 4284W

Micro Server Related Content

Videos

Product and Performance Information

open

1. Software and workloads used in performance tests may have been optimized for performance only on Intel® microprocessors. Performance tests, such as SYSmark* and MobileMark*, are measured using specific computer systems, components, software, operations, and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products.


2. Source: Performance comparison using SPECpower_ssj*2008 benchmark results using Java SE Runtime Environment (build 1.6.0_30-b12), Java HotSpot 64-Bit Server VM (build 20.5-b03, mixed mode). Baseline result of 3079 ssj*_ops/watt on prior generation Intel® Xeon ® Processor E3-1260L (Quad-Core, 2.4GHz, 8MB L3 cache), Enhanced Intel SpeedStep® Technology enabled, Intel® Turbo Boost Technology enabled, Intel® Hyper-Threading Technology (Intel® HT Technology) enabled, 8GB memory (2x 4GB DDR3-1333 UDIMM), 64GB SATA SSD, Windows Server2008R2 SP1. New result of 4291 ssj*_ops/watt on current generation Intel® Xeon® Processor E3-1265Lv2 (Quad-core, 2.5GHz, 8MB L3 cache, E0-stepping), Enhanced Intel SpeedStep® Technology enabled, Intel® Turbo Boost Technology enabled, Intel® Hyper-Threading Technology (Intel® HT Technology) enabled, 8GB memory (2x 4GB DDR3-1600 UDIMM), 64G 3Gb/s SATA SSD, Windows 2008 R2 SP1. Source: Intel internal testing as of Mar 2012.


3. Results have been estimated based on internal Intel analysis and are provided for informational purposes only. Any difference in system hardware or software design or configuration may affect actual performance.