June 2013: There's a lot of potentially confusing and contradictory information about blade energy efficiency out there – it's hard to know what exactly is being compared and how. Recently, Cisco seemed to claim that their blades were more energy efficient than HP. 1

We wanted to see if their claims would stand up to scrutiny and we discovered some interesting results.

Please consider the facts:

 

Fact 1: As a measure of power efficiency, performance per Watt is defined as rate of computation that can be delivered by a computer for every watt of power consumed. 2

Performance per Watt is the bottom line for comparisons across vendors that have different architectures. What really matters is how much work can a system do for a given electrical power load.

The Standard Performance Evaluation Corporation created the SPECpower_ssj2008 benchmark as an industry standard benchmark for comparing the performance per Watt (as expressed as overall ssj_ops/Watt) from different vendors. 3

 

Fact 2: Cisco's own power efficiency comparisons show that HP delivers better performance per Watt.

When Cisco tested fully configured blade enclosures from both Cisco and HP 4, Cisco's own test results showed that HP delivered more performance per Watt as follows:


A: Cisco UCS B200 M3 blade server
(single enclosure 8 servers/4 power supplies and NO included Fabric Interconnect):
3533 overall ssj_ops/Watt 5


B: Cisco UCS UCS B200 M3 blade server
(single enclosure 8 servers/4 power supplies servers WITH a pair of Fabric Interconnects):
2583 overall ssj_ops/Watt 6


C: HP BL460c Gen 8 blade server
(single enclosure 16 servers/ 6 power supplies):
3737 overall ssj_ops/Watt 7 = better performance/Watt than A or B above.


It's not every day that a competitor promotes benchmark results showing that your design delivers 6 to 45% better performance per Watt than theirs.

So when a vendor like Cisco tells the world that they agree that HP BladeSystem has better performance per Watt than Cisco UCS, that's significant news.

Improved power efficiency is only one aspect of the HP BladeSystem that can save customers 68% in total cost of ownership.8 To learn more about HP BladeSystem please visit: www.hp.com/go/bladesystem

 

Featured products

Analyst reports and white papers

1 Cisco: “Power Efficiency Comparison: Cisco UCS 5108 Blade Server Chassis and HP BladeSystem c7000 Enclosure”.
Results as of May 29, 2013
2 Performance per watt: http://en.wikipedia.org/wiki/Performance_per_watt
3 SPEC and the benchmark name SPECpower_ssj2008 are registered trademarks of the Standard Performance Evaluation Corporation (SPEC).
See http://www.spec.org/power_ssj2008/
4 Cisco: “Power Efficiency Comparison: Cisco UCS 5108 Blade Server Chassis and HP BladeSystem c7000 Enclosure” Hardware Configuration #2
5 Figure 11: Cisco: “Power Efficiency Comparison: Cisco UCS 5108 Blade Server Chassis and HP BladeSystem c7000 Enclosure
6 Figure 12: Cisco: “Power Efficiency Comparison: Cisco UCS 5108 Blade Server Chassis and HP BladeSystem c7000 Enclosure
7 Figure 13: Cisco: “Power Efficiency Comparison: Cisco UCS 5108 Blade Server Chassis and HP BladeSystem c7000 Enclosure
8 IDC white paper sponsored by HP, "Business Value of Blade Infrastructures", #227508R2, December 2012