IBM BladeCenter S beats the competition

IBM

Your business is not generic. IBM® understands that one size does not fit all. Our competition has yet to figure it out.

The BladeCenter® family of chassis forms your IT foundation and can be tailored to the range of applications, environments and performance requirements you have in your business today. Unlike other vendors with limited chassis and options, the BladeCenter platform has four chassis so you can pick the size, type and performance level you need. With a BladeCenter solution, most blades and switches can be easily moved among chassis, giving you incredibly flexible, mix-and-match deployment choices. And by supporting up to 14 blades per BladeCenter H or E chassis compared to the 8 to 10 from some competitors, the BladeCenter platform supports solutions with fewer chassis, switches, and cables to manage[1]. Also BladeCenter S is designed to work equally well in both the data center and the front office. It fits under a desk (using the Office Enablement Kit), is very quiet, and runs on standard 115V wall outlets.

HP

HP offers the c-Class and the c-Class Carrier-Grade chassis but delivers no compatibility between them. HP has separate switches and blades for each chassis type, requiring more part numbers and options for IT to manage. With IBM BladeCenter, most HS/LS/JS/PS/QS blade servers and switch modules ever released by IBM are supported in every BladeCenter chassis ever released by IBM, going back to 2002^[2].

Dell

Dell only offers one chassis offering, the PowerEdge M1000e chassis. Dell neglects the small and medium businesses For clients with less demanding I/O performance needs, IBM offers a choice of BladeCenter S or E chassis, which provide simple installations with smaller footprints and easy local storage. With Dell’s single chassis offering, they must try to oversell the M1000e or sell rack/tower servers.

Cisco

Not only does Cisco neglect choice, they also neglect to offer substantial blade density with their one chassis offering, the Cisco Unified Computing System (UCS) 5108 Blade Server Chassis. With IBM BladeCenter most chassis and components are interchangeable. If a user decides to move to a different BladeCenter chassis platform, they can take most internal components from the old chassis and install them into the new one – delivering extreme flexibility while often removing the need for significant additional capital expenditures when moving to a new chassis platform. Cisco’s one chassis offering effectively limits users to a single configuration. In addition to providing choice and flexibility, IBM BladeCenter provides a much denser environment. For instance, the BladeCenter H 9U chassis supports 14 blade servers compared to the Cisco 5108 6U chassis supporting only 8 blade servers.
With Cisco UCS, you would need to purchase more chassis – which means more power supplies, fans, and fabric extender I/O modules.

¹ A standard 42U rack enclosure can fit up to 84 IBM blades (168 sockets per rack using 2-socket servers) via 6 IBM BladeCenter E chassis measuring 7U and supporting 14 blade servers each.
A standard 42U rack enclosure can fit only up to 64 HP blades (128 sockets per rack using 2-socket servers) via 4 HP BladeSystem c7000 Enclosures measuring 10U and supporting 16 blade servers each.

IBM

IBM understands that you have a variety of applications — but that you still want to move to one, efficient, easy-to-manage environment. We offer up to four-socket Intel Xeon processor-based blades and IBM POWER® processor-based blades supporting a wide range of workloads on Microsoft Windows, multiple Linux distributions, IBM AIX®, IBM i, and Solaris 10 operating systems, you can select exactly the right servers for the workloads you need to support, even mixing-and-matching them in a single chassis.

In addition, IBM offers the following unique eX5 innovations to help you further reduce cost and complexity and pre-integrated systems and solutions that can help you achieve faster time to value.

• Two fully functioning 2-socket IBM BladeCenter®HX5 blade servers can be easily connected to form one 4-socket HX5 server, doubling all system resources or, the 4-socket HX5 blade can be physically separated back into two separate servers.
• FlexNode “no-touch” partitioning enables logical reconfiguring of a 4-socket HX5 blade into two HX5 2-socket blades without any physical system reconfiguration, so you get the freedom to run, for example, scale-out applications by day on two 2-processor systems and scale-up jobs by night on a 4-processor system.
• The MAX5 memory module allows you to add an additional 24 DIMMs to a base HX5 blade server without additional processors or software license costs.
• Preconfigured, pretested systems and solutions targeted to specific workloads can help your business achieve faster time to value.

• The HX5 optimized system for database speeds the deployment of transactional database workloads.
• The HX5 optimized system for virtualization is optimized to deliver more memory for more partitions and larger workloads.
• The IBM BladeCenter Foundation for Cloud provides a comprehensive converged solution that brings together servers, storage, networking, management, software and services needed to quickly establish a robust and virtualized environment. This highly reliable and scalable platform can easily be extended to a private cloud with the addition of IBM SmartCloud Entry software that includes leading edge self-service portal, automation and metering.
• The zEnterprise BladeCenter Extension (zBX) is an infrastructure for extending tried and true System z® quality of service and management capabilities across a set of integrated compute elements in the zEnterprise System.
• The i Edition Express for BladeCenter S is a complete integrated business system that offers an attractive alternative for existing small and midsized IBM i clients to reduce IT costs and complexity by consolidating their IBM i and Windows applications with an IBM BladeCenter solution.

HP

When it comes to blade solutions, HP lacks choice and technology innovation. For instance, with Virtual Fabric for IBM BladeCenter, IBM offers a range of I/O virtualization and convergence networking solutions based on advanced technologies from an ecosystem of Partners. If a customer has standardized on another vendor for switches, they will not have to rip-and-replace their existing hardware. HP requires Virtual Connect, an HP proprietary solution for their I/O virtualization solution. This means customers’ who have standardized on another vendor for switches will need to rip-and-replace their hardware.

In addition, IBM provides innovative tools and solutions to help customers get faster time to value. IBM FastSetup is a no-cost software tool that helps reduce the hands-on server setup time from days to minutes, particularly for larger deployments. IBM SmartCloud Entry provides a fully integrated software stack for transforming a virtualized environment to a cloud environment. And finally, IBM BladeCenter Foundation for Cloud, a comprehensive virtualization platform, is designed to easily scale to a robust, private cloud solution with the addition of IBM SmartCloud Entry software.

Dell

Dell doesn’t offer any comparable blade innovations or pre-integrated offerings.

Cisco

Cisco UCS only provides x86 blade servers. With IBM BladeCenter, you can truly reduce the complexity, management, and energy requirements of your IT operations by consolidating your x86, IBM i, IBM AIX and Linux workloads in a single bladed architecture. And though Cisco UCS delivers Vblock, a pre-integrated platform for virtualization and cloud computing, IBM delivers integrated and pretested solutions for cloud computing and virtualization with IBM BladeCenter Foundation for Cloud. In addition to providing a fully integrated virtualization platform, IBM BladeCenter Foundation for Cloud can easily be extended to a private cloud with the addition of IBM SmartCloud Entry software, and it can be done without having to rip-and-replace the installed infrastructure, unlike Cisco UCS Vblock.

IBM

IBM understands that building an enterprise infrastructure and/or virtual infrastructure requires end-to-end reliability. Moving to shared RAID storage can help increase your data and application availability and make management easier. In addition, IBM storage solutions enable many levels of flexibility that add capacity in a modular, nondisruptive manner. BladeCenter offerings include Direct Attach Storage (DAS), Network Attached Storage (NAS), as well as FCoE, Fibre Channel (FC), SAS, and iSCSI SAN connectivity using a range of drive types and RAID levels.

• Direct Attach Storage — The BladeCenter S chassis supports up to 12 SAS or SATA HDDs (or a combination of the two), accessible by any or all blade servers in the chassis.
• Network Attached Storage — IBM System Storage® N series provides heterogeneous storage access—using iSCSI and Fibre Channel—to a broad range of host and client systems.
• Storage Area Network — The innovative BladeCenter S chassis, offers an optional SAN within the chassis itself. For larger needs, BladeCenter S and other chassis can be partnered with the System Storage DS3000 entry storage family, which provides easy and affordable storage for SMB and large enterprise distributed environments running Microsoft Windows, Linux, and IBM AIX operating systems. For even greater storage needs, the DS5000 mid-range storage family is designed to deliver high-bandwidth performance to Windows, Linux, IBM AIX and IBM i environments. And with a modular design with models at multiple price-points, the DS5000 storage system family of products can be used as storage add-ons or integral components of multi-tiered enterprise infrastructures.

• iSCSI SAN — BladeCenter offers both a 1Gb hardware and a 1Gb and 10Gb software iSCSI initiator. Both support iSCSI boot. Works with the DS3000 and DS5000 storage families as well as other iSCSI-interface System Storage subsystems.
• SAS SAN – BladeCenter HX5 with SAS connectivity module can connect to DS3500 to create a high performance 6Gbps SAS environment.
• Fibre Channel SAN — BladeCenter integrates storage switches from vendors such as Brocade, Cisco and QLogic. And BladeCenter supports industry-leading Fibre Channel host bus adapters from Emulex and QLogic. These options work well with the IBM System Storage DS3000 and DS5000 series. IBM data tells us that more than 40% of BladeCenter users attach their blades to a SAN. This is likely due to the wide selection and cost savings associated with integrated Fibre Channel switch modules.
• FCoE – BladeCenter H and HT chassis provide support for FCoE switch modules that help consolidate the data and storage networks into a single, high performance, super-network.

If you're seeking local hard drives, IBM offers high performance/high-availability integrated storage:

Blades:

• Internal hot-swap or fixed solid-state drive — For increased reliability and lower power consumption within a blade.
• Internal hot-swap SAS drives — For I/O-intensive workloads within a blade.
• Internal high-IOPS SSD adapters — For extreme I/O-intensive workloads within a blade.
• Internal USB flash — For embedded hypervisor.
• Hot-swap SAS — For RAID-5 and using the BladeCenter Storage and I/O Expansion blade.

Chassis:

• Hot-swap SAS — For high performance workloads.
• Hot-swap SATA — For price/performance-optimized workloads. Available on BladeCenter S.
• Hot-swap SAS/SATA SANs — For easily configurable self-contained SANs.

HP

HP offers no compelling in-chassis, shared SAN solutions. In addition, HP lacks continuity across their storage platforms, which requires unnecessary use of IT resources.

Dell

Does Dell understand the power of blades? We think not. Dell has a one-size-fits-all mentality. This lack of vision is also evident in their storage offerings, which focuses on rack-server interoperability. In addition, Dell doesn’t offer SAS connectivity and overall, their storage products are low in performance compared to IBM System Storage Disk Systems.

Cisco

Cisco supports only FCoE from the chassis to the fabric interconnect.
Internal support for SAS, SATA, and SSD drives as well as a internal USB slot

IBM

The IBM BladeCenter platform is designed for maximum availability. IBM understands that resiliency is critical when consolidating multiple workloads into a single chassis. While other blade vendors talk about this concept, IBM has embraced it. The BladeCenter design incorporates rock-solid availability with many redundant features, to better protect you against a single, catastrophic fault taking down the entire system. With BladeCenter you get:

• Blade and chassis designed with no single point of failure:

• Dual power connections to each blade
• Dual I/O connections to each blade
• Dual paths through the backplane to I/O, power, and KVM
• Redundant N+N power bus
• Hot-swap/redundant HDDs and SSDs on selected blades

• BladeCenter Open Fabric Manager for automated failover of blades, I/O ports and network or SAN switches
• True N+N thermal solutions
• IBM light path diagnostics
• IBM Predictive Failure Analysis
• IBM First Failure Data Capture

Tools like Predictive Failure Analysis, light path diagnostics and First Failure Data Capture work together to help warn you about problems—even before they occur. That's the kind of thinking you need when everything is riding on your IT. If a problem does occur, BladeCenter Open Fabric Manager can automatically detect the failure and bring up a spare blade, or move traffic from a failing switch or I/O port to working ones.

HP

• Blade and chassis design with many single points of failure:

• Single power connections to each blade
• Single I/O connections on BL460c and BL465c blades
• Single I/O paths for mezzanine slots 2 and 3 on the BL490c and BL495c blades
• Single paths through the backplane to I/O, power and KVM
• All power connected to a single power bus

• N+1 thermal solutions
• No light path function on the individual blades
• No function comparable to IBM First Failure Data Capture
• HP’s virtual connect only works with their proprietary hardware.

Dell

• Blade and chassis design with many single points of failure:

• Single power connections to each blade
• Single I/O connections to each blade
• Single paths through the backplane to I/O, power and KVM
• All power connected to a single power bus

• N+1 thermal solutions
• No light path function on the individual blades
• No comparable function to IBM First Failure Data Capture

Cisco

• Blade and chassis design with many single points of failure:

• Single power connections to each blade
• Single I/O connections to each blade

• Cisco’s UCS Manager is embedded in the Fabric Interconnect switch and manages ONLY UCS products. Clients will require multiple interfaces even to manage other Cisco products. IBM Systems Director and BladeCenter Open Fabric Manager can help you manage a heterogeneous environment using one easy-to-use interface.

Can’t manage storage with UCS
Can’t deploy an OS or application with UCS Manager

IBM

Light path diagnostics is an IBM solution designed to increase availability and speed service by illuminating LEDs next to parts that need attention.

This revolutionary technology helps expedite repairs and minimize downtime by quickly and clearly identifying those components that need maintenance—even without power to the blade. Just push a button and the battery will highlight the problem. This is a major IBM serviceability advantage.

HP

HP Insight Display provides no "insight" if blades are removed from the chassis. There is no battery to power the lights, meaning it cannot indicate a failed part on a test bench.

Dell

Dell only offers a front-mounted retractable LCD that delivers basic, real-time-only information, but individual Dell blades do not have any light-path functions to help in identifying failed components.

Cisco

Currently, Cisco does not offer comparable technology.

IBM

First Failure Data Capture simplifies problem determination by creating detailed event logs through the Advanced Management Module. This is especially helpful when there are cascading issues, as problem logs can get overridden. First Failure Data Capture is designed to help you find the original event that caused the issue so corrective action can be taken.

HP

Currently, HP does not offer comparable technology.

Dell

Currently, Dell does not offer comparable technology.

Cisco

Cisco UCS blades support First failute data capture, however it’s currently limited to setting "thresholds" for things such as memory and QPI errors which would alert the user. This, however, is not automated.

IBM

With a commitment to interoperability and open standards, IBM provides a selection of high performance Ethernet switch modules from a choice of industry-leading vendors, including BNT, Brocade and Cisco. This means that if you have already standardized on switches from one vendor, you don’t have to rip-and-replace to implement a BladeCenter solution. And, if you haven’t standardized on one vendor, you have the flexibility to mix-and-match as needed.

BladeCenter offers these 10Gb Ethernet switches today:

• BNT 10Gb Virtual Fabric Switch Module
• Brocade Converged 10GbE Switch Module
• Cisco Nexus 4001I Switch Module
• IBM 10Gb Ethernet Pass Thru Module

We have certified and pretesting on 3rd party switching.

HP

HP offers no choice or flexibility with their proprietary HP-only offerings.

Dell

Dell offers only the PowerConnect M8024 10GbE Switch. One choice. Too little flexibility.

Cisco

Cisco UCS does not support any switching inside the chassis.

IBM

ONLY IBM monitors all vital system components for maximum availability. IBM Predictive Failure Analysis sends alerts in advance of potential hardware failure of key components, to trigger preemptive action, and IBM light path diagnostics leads a servicer directly to the failing component so it can be replaced quickly, potentially with no workload downtime.

*These components are not mentioned in HP's**, Dell's** or Cisco’s pre-failure warranty support information.

** Links reside outside of ibm.com

IBM

BladeCenter is designed from the ground up to dramatically improve power utilization and reduce energy costs. In fact, aggregate power savings of BladeCenter vs. 1U servers and related external equipment can be as much as 30-40%. This can be attributed to a well-planned system architecture that includes high-efficiency power supplies and fans, lower-draw processors and memory, and a super-smart power delivery solution.

With IBM BladeCenter you get:

• Up to 94% efficiency^[4] with power supplies that use less energy and produce less waste heat that needs to be cooled than rack and tower power supplies that average only 70-80% efficiency.
• A BladeCenter S chassis that plugs into standard 115V office power outlets. It doesn’t require a data center, making it ideal for SMBs and distributed offices.
• Dynamic fan speeds — In the event of a fan failure, the other fans will run faster to compensate until the failing fan is replaced. Competitive fans must run faster at all times, just in case, wasting power.
• Solid-state drives that consume up to 80% less energy than 2.5-inch HDDs and up to 88% less than 3.5-inch HDDs (while producing higher IOPS performance).
• The ability to add an integrated 10Gb Ethernet switch (BladeCenter H and HT), which can be up to 95% more energy efficient compared to external switch offerings with individual power supplies.
• A selection of low-voltage processors that that intelligently adjust their voltage and frequency.
• 1.5V DDR3 DIMMs that consume 10-15% less energy than the 1.8V DDR2 DIMMs used in older servers.
• 1.35V DDR3 DIMMs that consume up to 19% less energy than 1.5V DDR3 DIMMs.
• IBM Systems Director Active Energy Manager™, a tool that can help you plan, predict, monitor, and actively manage the energy consumption of your BladeCenter servers.

HP

Unlike HP, IBM has pioneered new technologies that allow IBM BladeCenter blades to generate less heat output and use less energy to cool the system. In fact, Edison Group, an independent technology analysis and consulting firm, found that it takes approximately 31% less airflow to cool the IBM BladeCenter E than an equivalently configured HP BladeSystem c7000.^[5] In addition, BladeCenter chassis also offer significantly better performance-per-watt than equivalent HP systems: up to a 12% advantage with IBM BladeCenter H and up to a 22% advantage with IBM BladeCenter E.

Dell

While Dell says they provide efficient power management tools, their dynamic power monitoring through Dell OpenManage IT Assistant sets the power cap through a hard threshold. IBM can use power trending to set the power cap using IBM Systems Director Active Energy Manager. Active Energy Manager enables power consumption to be monitored through IBM Systems Director, even when the BladeCenter system is powered down. In addition, Active Energy Manager works with IBM System x®, BladeCenter, iDataPlex™, and POWER servers, and also monitors storage and third-party equipment through the use of intelligent PDUs. Dell’s Power Manager can only manage Dell blades.

Cisco

Cisco’s UCS Manager offers no power or thermal monitoring capabilities and contains no ability to reduce power consumption via capping.

⁴ Based upon the September 2010 80 PLUS Verification and Testing Report (PDF, 98.8KB) where the IBM BladeCenter H chassis using 2980W power supplies earned a Platinum-level rating for energy efficiency and power factor. 80 PLUS is a certification program that promotes highly efficient power supplies (greater than 80% efficiency in the active mode) in technology applications.

⁵ Based upon the May 2010 whitepaper from Edison Group (PDF, 1.9MB)

IBM

The Brocade Converged 10GbE Switch Module offers one of the industry's best integrated I/O solutions. The compact design incorporates Ethernet and Fibre Channel switching and provides a total of 30 ports, including eight external 10Gb Ethernet CEE ports for LAN connectivity and eight external 8Gb Fibre Channel ports for storage and SAN connectivity. With the high integration of this module, you can achieve all of your networking and storage I/O needs with a single module.

HP

HP’s FlexFabric proprietary solution requires HP’s Virtual Connect Switch and expensive 10Gb infrastructure outside the chassis. In addition to providing rigid blade solutions, HP’s FCoE offerings provide half the bandwidth^[6] of the IBM Brocade Converged Switch. Simply put, HP FCoE solutions offer no choice, no flexibility, and much lower performance.

Dell

Just like HP’s offering, Dell’s FCoE switch falls short in performance delivering 22% less bandwidth overall and 50% less Fibre Channel bandwidth^[7] compared to the IBM Brocade Converged switch.

Cisco

Cisco UCS provides only four 10Gb Ethernet ports for each of the 8 blades servers in their 5108 Blade Server Chassis. Again, there is no integrated FCoE switching available with the Cisco UCS.
And because many clients like their already stable SAN implementation, and do not want to change it, if they implement UCS, they are forced to use a Cisco version of Fibre Channel over Ethernet (FCoE). You cannot implement Fibre Channel down to the server in a UCS construct. This is an issue for many clients. With BladeCenter and System x, clients are not forced a move to FCoE until they are ready. And once the client is ready, IBM supports FCoE solutions from multiple vendors, including BNT, Brocade, and Cisco — offering clients a choice versus vendor lock-in.

⁶ HP’s Flex-10 technology with FCoE solution provides only four 10Gb Ethernet uplink ports (80Gb per second bidirectional Ethernet throughput) and four 8Gb Fibre Channel uplink ports (64Gb per second bidirectional Fibre Channel throughput) for a total of 144Gb per second throughput. IBM’s Brocade Converged switch provides eight 10Gb Ethernet uplink ports (160Gb per second bidirectional Ethernet throughput) and eight 8Gb Fibre Channel uplink ports (128Gb per second bidirectional Fibre Channel throughput) for a total of 288Gb per second throughput, doubling the overall throughput performance. http://h18004.www1.hp.com/products/blades/components/ethernet/10gb24/index.html (link resides outside of ibm.com)

⁷ Dell’s FCoE solution provides eight 10Gb Ethernet uplink ports (160Gb per second bidirectional Ethernet throughput) and only four 8Gb Fibre Channel uplink ports (64Gb per second bidirectional Fibre Channel throughput) for a total of 224Gb per second throughput. IBM’s Brocade Converged switch provides eight 10Gb Ethernet uplink ports (160Gb per second bidirectional Ethernet throughput) and eight 8Gb Fibre Channel uplink ports (128Gb per second bidirectional Fibre Channel throughput) for a total of 288Gb per second throughput, doubling the Fibre Channel throughput performance. http://www.dell.com/us/business/p/dell-m8428-k/pd(link resides outside of ibm.com)

IBM

The deployment of server virtualization technologies in data centers requires significant efforts in providing sufficient network I/O bandwidth to satisfy the demand of virtualized applications and services. With a Virtual Fabric for IBM BladeCenter solution, you can significantly reduce complexity and cost by reducing switch, cable and adapter costs, while achieving better energy efficiency. At the same time, the virtual fabric solution gives you the capability to allocate and adjust bandwidth based on actual requirements and the flexibility to upgrade to hardware iSCSI (or FCoE—statement of direction), without losing your existing investment in switches and adapters.

Virtual Fabric for IBM BladeCenter supports:

• Multiple virtual ports and protocols — Ethernet, iSCSI, and FCoE from a single physical port while allowing shared bandwidth across multiple applications.
• True line-rate 10Gb performance
• Up to 8 virtual ports operating anywhere between 100 Mb and 10 Gb, running as Ethernet, iSCSI, or FCoE
• Intelligent Failure Monitoring that enables automatic failover between physical or virtual ports in the event of an uplink port failure.

HP

HP's Virtual Connect Flex-10 offering delivers similar functionality. However, IBM now supports vNIC2 which can work with Cisco Nexus 4001I, BNT and 10G Pass thru. It will also work with Brocade FCoE switch in the future.

Dell

Dell offers no comparable technology.

Cisco

Similar technology, but IBM Virtual Fabric offers flexible switch options while Cisco’s solution is proprietary
Virtual Fabric with virtual interface card and FEX technology.