Compare blade servers. Learn the benefits of IBM BladeCenter compared to competitive blade servers.
Tab navigation
- Overview- selected tab,
- Dare to compare
- Migration path
- Get the facts
All blades are not created equal.
IBM® BladeCenter® is a highly integrated system comprising of blade servers, storage, management and networking options that help reduce management complexity, increase performance and energy-efficiency while significantly reducing costs with a highly reliable and flexible design that deters the need for “rip and replace” as capacity needs change. Top 5 reasons to choose IBM BladeCenter:
1. The widest choice of compatible chassis, blade servers, storage and networking offerings
2. Rock-solid availability with multiple layers of redundancy
3. End-to-end reliability and a commitment to best practices
4. Excellent power efficiencies with pioneering cooling technologies
5. Superior I/O performance and switches based on open technology
Is your IT environment in need of better performance and efficiencies?
IBM BladeCenter HS22 and HS22V blade servers can help improve the economics of your data center with up to:
- 23X faster performance than older servers1
- 96% reduction in energy costs alone2
- 99% IT footprint reduction3
- 50% lower connectivity costs4
- 92% fewer networking cables5
IBM BladeCenter: smarter systems for a smarter planet
Today’s data center environment is tougher than ever. You need to reduce IT cost and complexity. You need to manage space requirements, power consumption and heat output. And you need to do these things while increasing flexibility, utilization and manageability. So how do you manage the explosion of data and still reduce costs? IBM BladeCenter has the flexibility and reliability you need to manage the massive – and growing – amounts of data that your systems are creating without adding complexity. The innovative, open BladeCenter design offers a true alternative to today's sprawling racks and overheated server rooms. Migrate to the blade server solution that lets you do it all and reduce costs. You have nothing to lose but complexity.
Integrate:
- Servers, networking, storage, & systems management
- Virtually any networking standard from multiple vendors
- End-to-end management
- Linux®, UNIX®/IBM AIX®, IBM i and Windows®
- 1-, 2- and 4-socket x86 and IBM POWER® processor-based blade servers
- Multiple workloads on a platform designed for virtualization
Simplify:
- Management with a single point of control for the entire data center
- Deployment with automated provisioning tools
- Qualification with scalable, flexible building blocks
- Networking with Virtual Fabric and FCoE
- Complexity with an open platform that fits right into your data center
Reduce:
- Costs with fewer parts to buy and lower operating overhead
- Power consumption with a high-efficiency system
- Space with high-density blades
- Cables with integrated switching
- Risk with no single point of failure via multicomponent PFA and automatic recovery
- Waste with a stable, proven platform that has stood the test of time
[1] Performance and energy savings calculations are based on actual SPEC results and an IBM estimate of SPECint_rate2006.
Rack to blade comparisons:
2010 HS22/HS22V to a 2005 xSeries 346 2U rack server SPECjbb2005 ratio = 23.1-to-1
- HS22/HS22V blade server configuration: 2S 6C Xeon X5680 (Intel Turbo Boost Technology up to 3.60GHz and 12MB cache) with 24GB (12 x 2GB) of DDR3 PC3-10600R memory, and 1 SSD; average power draw with chassis burden = 372W under load, SPECjbb2005 = 915,323 bops
- x346 2U rack server: 2S 2C Intel Xeon (2.80GHz 2MB cache) with 8GB (8 x 1GB) and 1 HDD; average power draw = 384W under load, SPECjbb2005 = 39,585 bops.
2010 HS22/HS22V to a 2006 System x3650 2U rack server SPECjbb2005 ratio = 7.95-to-1
- HS22/HS22V blade server configuration: 2S 6C Xeon X5680 (Intel Turbo Boost Technology up to 3.60GHz and 12MB cache) with 24GB (12 x 2GB) of DDR PC3-10600R memory, and 1 HDD; average power draw with chassis burden = 372W under load, SPECjbb2005 = 913,456 bops.
- x3650 2U rack server: 2S 2C Xeon 5160 (3.00GHz 4MB cache) with 8GB (8 x 1GB) and 1 HDD; average power draw: 324W under load, SPECjbb2005 = 114,941 bops.
Blade to blade comparisons:
2010 HS22/HS22V (SPECint_rate2006) vs. 2005 HS20 (IBM estimate of SPECint_rate2006) ratio = 18-to-1
- HS22/HS22V blade server configuration: 2S 6C Intel Xeon X5680 (Intel Turbo Boost Technology up to 3.60GHz and 12MB cache) with 24GB (12 x 2GB) of DDR3 PC3-10600R memory, and 1 SSD; total power with chassis burden = 372W under load, SPECint_rate2006 = 377
- HS20 blade server configuration: 2S 1C Xeon (3.80GHz 2MB cache) with 8GB (4 x 2GB) and 1 HDD; total power with chassis burden = 280W under load, IBM estimate of SPECint_rate2006 = 21
2010 HS22/HS22V vs. 2007 HS21 XM (SPECint_rate_base2006) ratio = 6.65-to-1
- HS22/HS22V blade server configuration: 2S 6C Intel Xeon X5680 (Intel Turbo Boost Technology up to 3.60GHz and 12MB cache) with 24GB (12 x 2GB) of DDR3 PC3-10600R memory, and 1 SSD; total power with chassis burden = 372W under load, SPECint_rate_base2006 = 352
- HS21 XM blade server configuration: 2S 2C Intel Xeon 5160 (3.00GHz) with 16GB (8 x 2GB) of DDR2-5300F ECC memory, and 1 HDD; total power with chassis burden = 319W under load, SPECint_rate_base2006 = 52.9
Source: www.spec.org (link resides outside of ibm.com). SPEC and SPECint are trademarks of the Standard Performance Evaluation Corporation (SPEC).
[2] Using the 22.1-to-1 performance ratio (of this particular processor) for today's HS22/HS22V blades to a 2005 xSeries 346 2U rack server, 1 BladeCenter H chassis containing 14 HS22V blades would provide the computing performance of approximately 310 2U rack servers. 1 BladeCenter H chassis containing 14 HS22V blade servers consumes an average of 4,382W. 310 2U rack servers consume 119,040W. Thus, the energy savings for replacing 310 x346 2U rack servers with 1 BladeCenter H chassis is 96.3%.
- HS22/HS22V blade server with "reduced voltage" processor configuration: 2S 6C Xeon X5670 (Intel Turbo Boost Technology up to 3.33GHz and 12MB cache) with 24GB (12 x 2GB) of DDR3 PC3-10600R memory, and 1 SSD; average power draw with chassis burden = 313W under load, SPECjbb2005 = 875,369 bops.
- x346 2U rack server: 2S 2C Intel Xeon (2.80GHz 2MB cache) with 8GB (8 x 1GB) and 1 HDD; average power draw = 384W under load, SPECjbb2005 = 39,585 bops.
[3] Data center floor space cost estimate used is $500 per square foot, averaging 8 square feet per standard enterprise 42U rack. IT footprint compares 1U per rack server vs. 0.5U per HS22/HS22V blade. 14 HS22/HS22V blades in 7U (.17 of one 42U rack)
BladeCenter E chassis provides equal or greater performance than 310 x346 2U rack servers inside 14.76 rack enclosures, for a savings of 14.59 racks or 99%. In terms of floor space that would save approximately 117 sq. ft. (613U).
(Realistically, in most data centers 2U rack servers cannot be populated that densely, due to air flow and power issues. So the IT footprint savings would be even greater for BladeCenter E.)
[4] Up to 50% lower connectivity costs based on IBM BladeCenter Open Fabric Manager vs. HP Virtual Connect web pricing as of 8/1/10
[5] Up to 92% fewer networking cables based on calculations of: one 42U rack populated with 4 BladeCenter H chassis populated with 14 blades each (56 total) supporting 12 I/O ports per blade (672 total ports) requiring a total of 64 cables per rack for 6 integrated switch modules (8 cables from the Virtual Fabric switches and 8 cables from other LAN switches; the VF switches support both LAN and SAN traffic, so separate SAN cabling is unnecessary). An equivalent 56 1U rack servers populated in 2 rack enclosures supporting 14 I/O ports per server (12 LAN cables and 2 SAN cables connected to corresponding external LAN and SAN switches) requiring a total of 784 cables. This doesn’t even include the 85% reduction in cabling for power cords (16 per rack for BCH vs. 112 for 56 1U servers with redundant power).

