Magic Quadrant for Blade Servers
The blade server market is becoming increasingly polarized, with vendors deploying high-end integrated systems to help increase margins and share of wallet, while the lower end of the market is inundated by multinode server designs that address the market's appetite for extreme scale-out workloads.
Within this evaluation of the blade server market, we also include other modular server form factors (such as multinode), given their proximity to the market in terms of functionality and user buying criteria.
A blade server is a modular compute platform that combines blades into a custom-designed chassis to create a fully functioning system. Multiple chassis may then combine within a rack to create a larger system, and multiple racks may be combined to create a large system that could consume a whole aisle or container. Because blades can be bought and deployed individually, they become an efficient way to add compute resources when maximum granularity is desired.
Although blades provide the computing power, the chassis is really the most crucial investment. The chassis provides power and cooling provisioning to all blades, along with various common management functions. The chassis supports a backplane that provides connectivity (and even aggregation) from server to server, or from server to storage or the network, but network and storage input/output (I/O) can also be directly routed to the blades. While blade servers can have onboard storage, the trend is toward diskless blades that are booted direct from a storage area network (SAN). While the blades will conform to a common form factor (within the vendor's own product strategy) to fit the chassis, they can have different processor and memory characteristics to address a variety of workloads, and specialized storage, graphics and other blades can be deployed together with servers. Some blade vendors can combine two or more blades to become a larger, logical computer. Blade chassis designs are also highly proprietary; there are no standards for chassis interoperability between manufacturers, so users are effectively locked into whichever chassis design they favor.
It is common for blades with higher complements of processors or storage to be wider, so that two or more chassis slots are consumed. There are four common form factors: full-height single width, full-height double width, half-height single width and half-height half width (for example, two blades in a single half-height slot). Blade chassis capacity can vary, and may be populated with blades of different types, including additional memory, storage devices and network switches, or other I/O modules for added connectivity. Most blade chassis are designed to fit within standard 19-inch racks, but some earlier enterprise blade platforms have been based on other dimensions. It is easy to regard blade servers as being a distinct form factor that addresses different market requirements compared with rack-optimized servers. But all blades, by definition, leverage a rack-based topology (usually based on the standard 19-inch form factor). With each generation, the distinction between blades and conventional rack-based servers becomes more blurred. The distinction is even harder to maintain with the advent of multinode (skinless) servers, which, like blades, utilize common system components, such as shared power supplies and cooling fans, and enable easy hardware provisioning. Unlike blades, multinode servers are usually deployed horizontally in trays that fit into the rack, but connectivity is equally proprietary. Blades become hybrid solutions that exploit the standardization of the 19-inch rack form factor, while imposing proprietary integration within the chassis.
Blades are not the only form of modular server; multinode servers are an even more rack-dense form factor that has emerged during the past four years to address many extreme scale-out workload requirements that, ironically, blades were first designed to address. We now see the emergence of extreme low-energy servers (typically based on low-cost processors from Intel, AMD and ARM) that will extend the horizontal scaling further — again, usually based on some modular, tray-based form factor.
True multinode servers typically lack the availability of the richer tooling that benefits blade environments, so the two form factors address distinctly different workload needs. As the addressable market for blade servers evolved toward more sophisticated and diverse workloads, a vacuum in the server market gradually formed and blades became overengineered for their original market objectives. Multinode servers are an alternative form of modular design developed to fill that vacuum. They are designed with a reduced number of racks, chassis and, in some cases, motherboard components to maximize server density potential and to reduce manufacturing cost, material use and power consumption. Typical designs involve a lack of outside sheet metal coverings (hence, the commonly used term "skinless"), compared with individual servers, as well as shared power and cooling resources within the rack frame.
At first glance, multinode servers share many common attributes with blade servers, which explains why some vendors regard the markets for blade servers and multinode servers as synonymous. Like blades, multinode servers are designed to slide into a common chassis, enabling the quick and easy addition of new components, and the replacement of failed components. They may rely on common components, such as power supplies, cooling fans and I/O, which become functions of the chassis, not the multinode server. They are usually based on a standard x86 architecture, run a regular Windows or Linux workload, and conform to the 19-inch-rack-width standard.
In addition, the emergence of extreme low-energy servers, based on Intel, AMD and ARM processors (and even potentially reduced instruction set computer [RISC] processors), will open the door to new software stacks. As with blades, the mounting technology for multinode servers will be dictated by the server manufacturer, and is proprietary. Workloads and situations that lend themselves well to multinode server approaches include applications that share server resources across a network, such as extreme analytics and high-performance computing (HPC) environments. Multinode servers offer an additional benefit: Because they use less material in the server infrastructure, less material needs to be replaced and/or recycled. Some blade vendors referenced in this research are actively marketing multinode server designs, either as an alternative to or as a variant of blade server strategies.
The blade server market has largely stalled since 2011, as cannibalization from multinode servers has eroded the growth that blades had gained through increased market adoption of fabric-based infrastructure (FBI). Based on 2012 data, blades still only represent about 12% of the total server market in units (down from 13% at the end of 2011) and 21% in revenue. This has barely changed from 2011; in comparison, multinode servers now outsell blade servers in unit terms (despite being sold for only a few years and addressing a more limited volume market). In 2012, multinode servers represented a 15% unit share (nearly doubled from 7.8% in 2011) and 10% revenue share (again, nearly doubled from 5.3% in 2011).
Because they favor more horizontally challenging workloads, most blade and multinode deployments favor x86 architectures; however, vendors such as HP, IBM and Oracle ship non-x86 blades, primarily targeted at Unix users, and vendors such as Supermicro, SeaMicro (now owned by AMD), Dell, Huawei and HP are bringing low-power Intel-, AMD- and/or ARM-based servers to market.
Because multinode servers are taking over the low-end workload space traditionally occupied by blade servers, and because blade servers are usually the foundation for vendor FBI strategies, we are seeing more adoption of blades in production environments for complex applications — such as high-end database serving, data warehousing, ERP and CRM. Blade servers are also well-suited for complex in-memory database workloads. This has led to an increasing technology overlap among blade, multinode and rack-optimized servers, thus driving a need for vendors with a presence in all three form factors to be more transparent about workload optimization for each competing form factor. We recommend that customers continue to demand valid references and proof points for all workload scenarios that push the upper (or lower) boundaries of viable blade server implementation.
After a long period in which HP and IBM dominated the blade server market, it is becoming slightly more even with the 2009 entry of Cisco. Despite being a newcomer to the server market, Cisco is carving out a strong share of the blade server market that is rivaling IBM in many geographies (and even rivaling HP in a few geographic markets). HP, IBM and Cisco now control 80% of the blade server market (based on 2012 shipments). However, HP remains the volume market leader by far, with more than the revenue share of IBM and Cisco combined. As blade servers form the foundation of most vendor FBI strategies, the market will continue to attract investments from most server vendors, especially because Gartner believes that FBI products have much higher profit margins and vendor lock-in than other x86 server businesses. The market potential for private and public cloud server infrastructures also provides a natural opportunity for blade (and multinode) servers, as most cloud infrastructures are likely to be based on highly virtualized x86 platforms that are well-suited to the need for easy and frequent hardware provisioning.
Source: Gartner (April 2013)
Bull offers several blade or multinode solutions that address a variety of market needs, with a special focus on the HPC market segment. The NovaScale x86 blade family is targeted at business computing needs, while Bull's bullx server design is aimed at HPC, analytics and other extreme scaling requirements. To extend its international reach beyond the Western European heartlands, Bull has collaborative ventures with regionalized vendors and fulfillment partners in multiple countries, including emerging geographies. With just under 5% global market share, it is a viable blade server vendor, subject to the limited geographic reach that becomes a gating factor to execution ability.
- Bull maintains a viable presence in Western Europe, in addition to vertical niches across multiple geographies in industries such as financial services and the public sector.
- In recent years, Bull has become a well-established HPC market contender with a strong (and rapidly growing) international presence.
- The vendor has entered into several OEM and other fulfillment agreements to expand its technology reach in emerging markets.
- Bull is committed to technology innovation, especially energy efficiency and the management of large-scale clusters.
- Bull's limited regional presence is frequently a gating factor for its potential as a partner for multinational implementations.
- The vendor has a limited channel presence, particularly outside Western Europe, where it is dependent on OEM/co-marketing relationships.
- Bull will have to compete for cloud infrastructure provider priority against significantly larger global vendors in a finite service provider community.
Despite its relative youth as a mainstream server vendor, Cisco has established itself as a recognized Leader in the blade server space. The Unified Computing System (UCS) is highly innovative, and is targeted at highly integrated and virtualized enterprise requirements, with a growing focus on service providers. During 2012, Cisco launched additional initiatives to price and position UCS as a viable midmarket solution, along with management console improvements. Continuing management tool vendor acquisitions (such as Cloupia) are helping to reinforce market confidence in Cisco's commitment to the server market. The UCS architecture differs from that of most other vendors by deploying a top-of-rack switch as a fabric director, which is able to assign virtual personalities to the blades that become automatically provisioned on installation. Most other vendors need to deploy a subnetwork to achieve a similar result. UCS, therefore, becomes a form of integrated system where part of the total platform is networking-based, rather than computing-based. With no proprietary storage offering, UCS does not fulfill all the requirements for a complete FBI, but Cisco's design allows users to integrate and leverage their existing storage. The UCS portfolio is not limited to blades; Cisco sells a limited range of rack-optimized servers, and, in 2012, it became the first vendor to offer common identity profiles that function across both blade and rack servers, thus eliminating management barriers and variations between form factors. To date, Cisco has made very little impact on the rack-optimized server market, as its marketing efforts have concentrated on achieving early success in the blade market.
With no in-house storage strategy, Cisco must partner with various vendors to create predefined integrated system solutions based on UCS; notable examples are VCE Vblock (EMC-based), Cisco/NetApp FlexPod (NetApp-based) and a recent alliance with Hitachi. Cisco also has developing relationships with independent software vendors (ISVs), such as the Virtualization Experience Infrastructure (VXI) initiative between it and Citrix for infrastructure to support hosted virtual desktops. As a result of steady market progress through 2012 (which, in a static blade market, has been achieved to the detriment of rival vendors), Cisco has further consolidated its position as a Leader in this Magic Quadrant, with 14% global market share, although it can be difficult to validate the precise share between Cisco UCS and solutions such as VCE Vblock and Cisco/NetApp FlexPod that leverage UCS technology.
- Cisco is a global corporation with a recognized and respected presence in most data centers, due to its strong market share in networking.
- UCS is an enterprise-class platform with good integration of networking, virtualization and management tools, and is strengthened by ongoing acquisitions that enhance management capability.
- Cisco has extended the UCS portfolio during 2012 to better address midmarket and small or midsize business opportunities.
- Vblock and FlexPod are proven FBI solutions that provide Cisco with cross-selling opportunities to the broader, combined installed bases of partner organizations; further relationships with vendors such as Hitachi and Citrix are being developed.
- The 2012 acquisition of Cloupia should augment investments to strengthen Cisco's management software portfolio.
- Despite the rapid growth of its blade server business, Cisco's rack-optimized server market share significantly trails that of the established major vendors. This limits its opportunity to be recognized as a leading vendor across the whole server market.
- Cisco's strategy remains dependent on alliances with storage vendors (and management tool vendors, in some situations) to create a complete offering.
- The perception of Gartner clients is that Cisco and EMC's close natural partnership appears to be increasingly strained, with both vendors investing in partnerships and acquisitions that compete with the other.
- Strategic alliances with key OS and application vendors are still maturing for Cisco as a relative newcomer to the server market.
- With no presence in the market for multinode servers, Cisco is exposed to the rapid growth of this sector, which is taking business away from low-end blades.
Dell offers Intel Xeon and AMD Opteron blade servers that are well-engineered, enterprise-class platforms that fit well with the rest of Dell's x86 server portfolio. Dell's blade servers target a broad range of market needs and geographies. The vendor has focused its recent energies on expanding the scope of its entire FBI strategy, in which the M-Series blade family plays a key role. The announcement of Dell's Active System strategy brings together blade server technology with recent hardware acquisitions, such as Force10 Networks, RNA Networks and Compellent, along with software assets from Quest Software, Scalent and Gale Technologies. Existing FBI strategies such as vStart and Virtual Integrated System (VIS) are being further incorporated into the Active System strategy during 2013.
Meanwhile, the vendor comfortably maintains the market leadership position in the growing market for multinode servers with the PowerEdge C line, and its Data Center Solutions (DCS) division has been created to target cloud services providers and other buying centers for extreme scaling with customized designs. These innovations put Dell on a trajectory to create greater market disruption, and initially helped the vendor defend its market share. Dell maintains a viable blade market presence (now at just under 7% global market share), but has continued to gradually lose share to leading rivals.
- As a mainstream, x86 server market leader, Dell has extensive cross-selling opportunities to a large and growing installed base that spans multinode, blades, racks and towers.
- The acquisitions of Compellent, Quest Software, Force10 Networks, Gale Technologies and RNA Networks strengthen Dell's fabric computing message.
- The vendor is the clear leader in the rapidly growing market for x86-based multinode extreme scale-out servers.
- Dell has an aggressive pricing policy and a strong midmarket presence.
- It has focused innovation in areas such as memory aggregation, general-purpose graphics processing unit (GPGPU) support, cooling and virtual I/O.
- Dell has continued to lose blade market share during the past two years, and is now the No. 4 vendor (at best) in most geographic markets.
- Dell's data center strategy is the best it has ever been; but despite its No. 1 or No. 2 position in most x86 market segments, credibility among most enterprise buying centers remains very hard for the vendor to win.
- Its FBI messaging has been confused by overlapping acquisitions, partnerships and branding policy; Dell must achieve messaging around the new Active System branding to fully leverage recent acquisitions and establish a viable FBI presence.
Fujitsu offers a broad range of blade server and multinode server offerings, although with only 1.5% global blade market share; its presence is stronger in other server markets. This includes the high-end Primergy BX900 Dynamic Cube platform, the Primergy BX400 for midsize businesses and the Primergy CX multinode server design targeted at HPC, cloud and other rack-dense requirements. In addition to its own server management suite, a collaborative agreement with Egenera is enabling Fujitsu to grow modest business in North America and EMEA. Fujitsu is executing better on an international scale, but messaging remains weak, resulting in a Challenger position in this Magic Quadrant.
- Fujitsu blade and multinode server designs are highly innovative and compete well with designs from better-known rivals.
- A strong portfolio of rack-optimized, blade, tower and multinode designs gives Fujitsu extensive cross-selling opportunities to a large and growing installed base.
- The vendor has strong vertical market strength (especially in the public sector), and a strong regional presence in Western Europe (particularly Germany and the U.K.) and Japan.
- Fujitsu has a limited track record as a volume supplier outside Japan and Western Europe.
- The vendor is ramping up its multinode business, but recognition remains low for its capabilities in topical modular server markets, such as FBI and multinode servers, mainly due to limited market awareness and passive marketing.
- Fujitsu has a limited channel presence, especially in North America.
Hitachi has made quiet, but steady progress with its blade server strategy during 2012, although geographic implementation remains firmly skewed toward the domestic Japanese market and global market share is just over 1.5%. However, Hitachi is learning the "adjacency market" technique that Cisco employed to penetrate the blade server market as a natural extension to its network market presence. In Hitachi's case, this policy is based on the strong storage market presence it has achieved in most markets. Hitachi blades are owned by the Hitachi Data Systems brand, and address a broad set of enterprise and midmarket requirements. The Unified Compute Platform integrates Hitachi's Compute Blades and storage technology to create an integrated system aimed at service providers and enterprises. Hitachi innovates strongly around blade aggregation and highly integrated virtualization. Gartner client inquiries consistently indicate that Hitachi's blade servers are highly popular among the vendor's installed base, and this has helped Hitachi to make gains in this latest Magic Quadrant update. During 2012, Hitachi Data Systems expanded its channel partner program to increase market penetration, particularly in the Americas and EMEA.
- Hitachi has a well-proven platform, with a strong Japanese installed base.
- The vendor offers chassis options that address enterprise and workgroup/departmental/branch requirements, as well as the Unified Compute Platform's integrated system.
- Hitachi has a very strong reputation as a storage vendor that it is leveraging through channel program expansion in multiple geographic markets.
- The vendor is committed to technology innovation, particularly in I/O and memory aggregation, as well as hardware-embedded virtualization.
- Support for Peripheral Component Interconnect Express (PCIe) ExpressModules provides Hitachi blades with very versatile I/O manageability.
- Feedback from Gartner clients on the Unified Compute Platform is very positive.
- Local sales and marketing execution in Western markets remains generally geared more toward Hitachi's storage business than servers.
- Hitachi remains relatively unknown as a server vendor outside Japan.
- Its channel presence is particularly limited, especially in EMEA and North America.
HP is maintaining volume blade market leadership, with a market share roughly equal to the combined share of the next four vendors (IBM, Cisco, Dell and Oracle). However, in a static market, HP's financial supremacy continues to be slowly chipped away at by rival vendors, with a 4% decline (from 49% to 45%) since 2010.
In terms of its blade server portfolio, the breadth of HP's available technology remains very high — the most extensive portfolio in the industry. This includes highly scalable individual blades such as the recently introduced four-socket BL660 (which is well-suited to processor/memory-intensive workloads like SAP Hana); the range of VirtualSystem, CloudSystem and AppSystem branded integrated systems (although not all AppSystem variants are blade-based); and the recent introduction of the BladeSystem C7000 and C3000 Platinum chassis, which offers significant performance gains over its predecessor. Most importantly, the new Platinum chassis claims complete forward/backward compatibility with any technology designed for the C-Class chassis.
In the short to medium term, the greatest expansion in addressable market opportunity will come from HP's multinode server investments. HP has launched new ProLiant SL designs targeted at workloads such as Hadoop and Vertica structured analytics. HP is also developing extreme scale-out computing solutions on the brand-new Moonshot server family, which extends the previous ProLiant SL potential into true hyperscale territory, by supporting up to 180 servers in a 4 rack unit (4U) chassis (45 servers at launch). HP's plan of record consists of interchangeable cartridge options, including multiple processors (Intel, ARM and AMD) and various specialty engines for graphics, security, etc., all bound by an internal switched fabric. In the short term, the HP Moonshot System servers will address a limited audience of large-scale cloud services providers and similar types of markets, but the concept opens up many new workload opportunities across multiple industries and governmental applications. In addition to Moonshot, HP's Project Odyssey will bring very large-scale, blade-based servers, based on Intel Xeon, that will help address the growing market for Unix migration and mission-critical x86 workloads.
Consistent articulation of its message across a broad channel is HP's greatest challenge. If the vendor cannot always outline the best strategy to deploy the range of blade and multinode servers it already has, then the introduction of the new Moonshot hyperscale form factor will only exacerbate the situation. Blade and multinode server buyers should shortlist HP on the basis of market achievement and reach, along with proven technology innovation. However, they should also demand clear positioning of the relative strengths and business arguments of each architectural approach.
- HP's blade market share is more than double the combined share of its two closest competitors.
- HP is further building on strong investment in management tools that already enable a single point of management across physical and virtual infrastructures (and across blade- and rack-based servers), with plans to extend this to common tools across compute, storage and networking.
- HP's blade and multinode chassis options address diverse market requirements — data center, workgroup/departmental/branch, mission-critical availability, extreme scale-out and hyperscale.
- The vendor is committed to blade innovation, particularly around its Virtual Connect and FlexFabric virtual I/O solutions, cooling, infrastructure autoprovisioning, blade aggregation and fabric-enabled infrastructure convergence.
- HP's launch of the Moonshot servers (and the pending launch of Odyssey servers) builds on the BladeSystem architecture to create the most comprehensive modular server portfolio in the industry.
- HP has made limited progress in 2012 with the positioning of such an extensive portfolio of server options, spanning everything from tower, rack, blade and multinode server form factors and value propositions.
- While Virtual Connect is well-established on blades, the lack of support for rack-based servers creates operational complexity for users who have both form factors, and inhibits capabilities such as live migration and high availability.
- Positioning of HP's generic BladeSystem servers alongside various integrated system designs (and the positioning of one integrated system against another) is a particular challenge, especially because two-thirds of blade server revenue is driven by the channel and it is relatively easy to create an integrated system experience using separate components.
- HP's Unix revenue has continued to decline through 2012, and this places increasing scrutiny on the blade-based Odyssey strategy, which should see products brought to market later in 2013.
- Early success for the much vaunted Moonshot strategy will be highly dependent on the ability to attract early adopters, and a viable software ecosystem that could be fragmented across different ARM variants.
During 2012, Huawei has continued to make steady progress with its blade and multinode server strategy. Although business is still heavily skewed toward the domestic Chinese market, the vendor is targeting new business in other Asia/Pacific countries and other fast-growing IT markets such as India, the Middle East, Latin America and Eastern Europe. Huawei plans aggressive expansion of its non-Chinese business in 2013. It has achieved 1% total market share (more than three times its share at the end of 2011). Although progress is slower, Huawei is also creating campaigns to achieve recognition and trust in North America and Western Europe. The vendor offers three modular server solutions. The E6000 is a full-height enterprise blade platform based on Intel Xeon that is mainly aimed at installed base users. In 2012, Huawei announced its new-generation blade server, the E9000, targeting a range of markets, including private cloud and HPC. The E9000 supports multiple two-socket and four-socket compute nodes and flexible switch modules, including 10 Gigabit Ethernet (GbE), FCoE and Infiniband. FusionCube is an integrated system based on the E9000 that is well-suited to cloud infrastructure rollouts and other highly virtualized instances. The X6000 is a multinode server design aimed at scale-out workloads (and Huawei plans to launch an even more extreme scale-out platform later in 2013).
Huawei is deploying a classic adjacency strategy, where it seeks to sell servers (and other data center technology) as a natural complement to its much-better-known core networking and telecommunications businesses. Huawei has particular strength in the telecommunications industry, and is now targeting other vertical industries, such as financial services and energy. The vendor has earned a reputation for strong support, and a willingness to make large investments in markets for longer-term gains.
- Huawei can leverage a very strong regional presence in China and across many Asia/Pacific countries.
- The vendor has good cross-selling potential, especially where it is a preferred telecommunications or switch/router vendor, and where Chinese and other Asian companies are expanding operations into Western markets.
- With a rapidly growing (but still small) blade market share, Huawei has the company breadth to build market awareness on a global scale.
- Huawei supports up to two full-height PCIe ExpressModules per blade, providing very versatile I/O manageability.
- The vendor is committed to delivering innovation around integrated systems, virtualization and extreme scale-out solutions.
- Despite the marketing communications campaigns under way, the Huawei corporate brand still has relatively little market awareness in Western markets.
- Huawei has struggled to overcome nationalistic sentiment and mistrust among North American buyers, particularly those in public-sector organizations.
- Recognition of Huawei as a server vendor is still limited in many regions (with the obvious exception of China).
The 2012 launch of Flex System and the PureSystems family was widely anticipated, and was vitally needed by IBM. Despite continuing innovations, the vendor's outgoing BladeCenter range steadily lost blade market share during 2010 and 2011. IBM market share lost four percentage points between 2010 and 2012 (25% to 21%), a proportionally larger drop than that of HP. However, in 4Q12, IBM recorded a 0.6% increase, which indicates that the ramp up of the new PureFlex System and Flex System products are more than offsetting the decline in BladeCenter business. In recent months, Gartner has seen a marked increase in client interest, and this will potentially manifest itself in a greater degree of market share recovery during 2013.
Flex System is a new-generation blade platform that, like BladeCenter before it, is able to host both Power and x86-based workloads, but is additionally engineered to enable strong integration of storage, networking and management. It is well-suited to support a variety of enterprise workload needs. Flex System then forms the foundation of PureFlex, IBM's integrated system that further integrates storage and networking assets, and the PureApplication integrated stack system. The PureApplication System integrates both DB2 and WebSphere, enabling PureApplication to address both single and multiple workloads. IBM has also launched the PureData System, which is targeted at big data workloads.
The BladeCenter range is still available, for both current users who need to extend current infrastructure and for more specialized needs, like extreme scale-out and Network Equipment-Building System (NEBS)-compliance. Users should validate the road map for further product development as part of their due diligence when making major long-term investments.
The Flex System launch has enabled IBM to make gains in this Magic Quadrant update, although market adoption is still ramping up and users should ensure that IBM has the presales and postsales support assets in their geographies, as well as appropriate references to prove both market readiness and suitability for purpose.
- IBM has a very broad set of blade chassis options that address multiple enterprise and workgroup/departmental/branch requirements, as well as more specialized needs, such as NEBS compliance.
- The vendor supports x86 and Power-based blades, spanning Flex System, BladeCenter and the zBX blade-based infrastructure within the System z architecture.
- IBM's blade strategy benefits from the vendor's extensive portfolio of end-to-end management tools that extend beyond servers to storage, and are augmented by the 2012 introduction of patterns that are designed to facilitate deployment of workload-optimized stacks.
- Because Flex System is a new chassis technology, it should remain viable and strategic for IBM through the decade.
- The PureFlex, PureApplication and PureData strategies provide IBM with the opportunity for highly verticalized platforms that integrate the hardware and software stack for multiple workloads.
- IBM has gradually lost blade server market share through the past three years — this share declined from 25% in 2010 to 21% in 2012. However, there is early evidence that Flex System and its derivatives are now returning the vendor to net growth.
- IBM's x86 server strategy has long lived in the shadow of much stronger marketing and messaging of Power and mainframe systems; the strong corporate commitment to PureSystems has yet to dispel speculation about the vendor's commitment to the x86 server market; and fresh speculation about IBM potentially divesting some or all of its x86 server business will create additional flux until the situation is resolved.
- Although we know IBM is investing strongly in reference and case study generation, client feedback indicates that the number of proven references remains limited.
- Despite the promise of PureApplication, IBM's marketing around integrated stack systems has not yet overcome the strength of Oracle's engineered system market momentum.
- Vendors such as Dell, Huawei, HP and Supermicro have more visible traction in the market for x86-based big data solutions. As IBM ramps up its PureData solution, this will potentially enable the company to redress the balance.
Oracle sells a family of x86 and SPARC-based blade servers built around the Sun Blade 6000 chassis that was originally brought to market by Sun Microsystems. Oracle does not heavily market these systems for volume-based new business, partly because it is focusing its x86 hardware energies primarily on the increasingly successful Engineered Systems (such as Exadata and Exalogic). The Engineered Systems are not blade-based platforms, and help to demonstrate that a viable, integrated system strategy does not have to be built on the use of blades. However, the success of Oracle's Engineered Systems strategy continues to drive market speculation about the strength of commitment that Oracle has to the general-purpose x86 server market. Because most Engineered Systems have Oracle x86 servers as the principal building block components, Oracle is likely to continue its investment in the x86 business.
Despite the lack of strong promotion, Oracle maintains a comfortable No. 5 market position in the blade server segment, with just over 3% global market share; however, this has declined from 5.6% in 2010. While demand for Oracle Sun Blades remains relatively high, the primary demand comes from the telecom sector, where Sun's long commitment to NEBS compliance has made Oracle a strategic vendor. Oracle is also committed to delivering total cost of ownership (TCO) advantages through simplified pricing and support, particularly for users of Oracle software products.
- Oracle is investing in Sun Blade technology and has a broad range of blade offerings, including Intel x86 and the latest-generation SPARC T5 processor that has just been introduced.
- NEBS compliance makes Oracle a strategic vendor for telecommunications vendors.
- Support for PCIe ExpressModules provides Oracle blades with very versatile I/O manageability, including the ability of every blade to have its own unique I/O configuration.
- Oracle is committed to technology innovation, particularly around energy management, hardware resilience, virtualization and management tools, as well as flash memory integration.
- Oracle's blade market share has gradually declined year over year, despite company initiatives to rebuild client and channel confidence in the technology gained from the Sun Microsystems acquisition.
- Passive and inconsistent messaging means that Oracle has failed to dispel market speculation that volume, general-purpose computers are not strategic to its future.
- With no presence in the market for multinode servers, Oracle is exposed to the rapid growth of this sector, which is taking business away from low-end blades.
With just over $62 million in 2012 revenue, SGI has the smallest share among the eligible server vendors, with 0.7% market share. However, the vendor has good recognition across many markets, with well-established international business distribution helped by the 2009 acquisition of SGI by Rackable Systems. That acquisition provided SGI with a comprehensive portfolio of server platforms that help it address a diverse set of workloads. Its strategic blade offerings are the SGI ICE X and SGI ICE 8400 blade-based clusters, both of which are leveraged from SGI-originated technology. These primarily address the HPC market, where the vendor has a long and proven track record. Meanwhile, the SGI CloudRack C2 and X2 are multinode servers that build on the traditional scale-out strengths of Rackable Systems to address the growing demand for big data workloads. SGI has refreshed HPC and CloudRack lines in recent months.
By recognizing that the Silicon Graphics name had better global brand recognition (the Rackable Systems brand was little known outside North America), the "new" SGI has been able to leverage both installed bases, albeit with the challenge of educating the market that the SGI brand now has meaning for more than just HPC users. This is an ongoing effort, particularly outside North America.
- SGI is one of the most established and recognized HPC leaders, with a significant global installed base across many vertical markets.
- The polarized nature of the traditional SGI and Rackable Systems businesses gives the vendor cross-selling opportunities in many segments.
- SGI is committed to technology innovation, particularly around extreme vertical and horizontal scaling, and integration of graphics accelerators.
- SGI has a complex set of blade and multinode server options that need comprehensive positioning to avoid confusion.
- Integrating two corporate product lines has resulted in many legacy Silicon Graphics and Rackable Systems users having nonstrategic technology that will need migration to the streamlined product portfolio that SGI now sells.
- The SGI brand continues to be mainly associated with HPC, particularly outside North America, where Rackable Systems only ever gained limited recognition.
We review and adjust our inclusion criteria for Magic Quadrants and MarketScopes as markets change. As a result of these adjustments, the mix of vendors in any Magic Quadrant or MarketScope may change over time. A vendor appearing in a Magic Quadrant or MarketScope one year and not the next does not necessarily indicate that we have changed our opinion of that vendor. This may be a reflection of a change in the market and, therefore, changed evaluation criteria, or a change of focus by a vendor.
No vendors were added for the 2013 update of this Magic Quadrant.
NEC was dropped from this Magic Quadrant. Its blade server sales remain strong (in sixth place worldwide), but the vendor is very oriented toward the domestic Japanese market.
Both blades and multinode (skinless) servers constitute a segment of the overall server market that is defined by its modular deployment, and most (but not all) server vendors invest in one or more modular server technologies. The main catalysts for inclusion in this Magic Quadrant are that the vendor should have had sales volume of at least $50 million globally during 2012, of which at least 5% should have come from beyond the home market for the vendor. This means obvious inclusion for the three vendors (HP, IBM and Cisco) that represent the majority of blade shipments worldwide, plus seven vendors that have a strong commitment to the market, albeit sometimes in niche geographic or vertical deployments. A small number of blade vendors were excluded either because their overall sales were less than $50 million, or their market presence is geographically extremely narrow (with less than 5% from international sales), or because they are legacy vendors that mainly address an installed base market where there is little or no new business that Gartner can evaluate.
Blade market execution is achieved through a combination of three methods:
- Large, established vendors with a strong installed base of rack-optimized servers are in a natural position to advocate the use of blades (and/or multinode servers) as a mainstream evolution. HP, IBM and Dell are good examples.
- Equally well-known vendors that have strong data center awareness (but in a different hardware discipline) can use the blade server market as an adjacency play to extend data center reach. Cisco, Hitachi and Huawei are good examples of this policy in action.
- Smaller vendors (plus larger vendors that are more specialized by geography or vertical industry) can leverage the advantages of blades for certain workload requirements, where they can maintain a viable presence in a more niche-oriented market. Vendors like Fujitsu, Bull and SGI fall into this category.
This creates a polarized market reflected in the Magic Quadrant, where every vendor is pursuing a blade server strategy that yields profitable business. For the larger vendors, blades introduce a new positioning challenge that can impact execution effectiveness, while more niche-oriented vendors must work to evolve their target markets and maintain added value.
Source: Gartner (April 2013)
Data center innovation is leading to a gradual convergence of technology classes that will become steadily more granular and component-based, and blades are a natural steppingstone (although not the only one) toward this state. Most vendor FBI strategies typically use blades as the foundation for their work (although viable exceptions exist, such as Oracle's Engineered Systems, which deliver all the attributes of an FBI, but are based on rack-optimized technology). Blade servers create new challenges for I/O and networking, which, in turn, puts an additional onus on the functionality and close integration of server management tools. This favors vendors that are leaders in this field, or that have strong management tool integration with third-party partners.
With each new processor generation, blades are increasingly capable of addressing challenging workloads. x86 blade servers with up to four sockets and 40 cores are becoming viable, and even larger x86 blade configurations are planned for the near future. Some vendors are already selling non-x86 blade platforms with the ability to scale to 32 sockets. However, while there are few workloads that demand such a degree of vertical scaling, blade server vendors are increasingly positioning the whole chassis — or even the whole rack or aisle — as a single system. The notion of a single system will be further driven by vendor strategies to deliver processor and/or memory aggregation across multiple blades; this will enable even larger and more complex workloads to be addressed, still with the advantage of a fine-grained technology addition or replacement. By "aggregation," we mean the logical and scalable integration of multiple components, such as CPU and memory. Meanwhile, multinode server designs offer some vendors an even denser approach that suits extreme scaling workload requirements. Where absolute maximum throughput, at the lowest cost of deployment and with minimum demand for sophisticated management tools or hardware resilience, is king, even blade servers can be overengineered for the task.
Source: Gartner (April 2013)
Blade market Leaders typically need to have built an enduring track record across multiple geographies, vertical markets and workload scenarios, or need to make a very strong and sustained market entry. For many years, the blade server market was dominated by two vendors — HP and IBM. Those two vendors are still volume leaders, but their degree of market leadership has been eroded since Cisco penetrated the market in 2009. In most geographies, the combined share of HP, Cisco and IBM revenues is 80%-plus. As blade market growth has dwindled (due mainly to the growth of multinode servers), the polarized nature of the market presents a challenge to other vendors seeking significant volume growth, as sustained achievement can only come from the partial decline of one or all the Leaders.
Challengers are likely to be capable vendors that are focusing their blade strategies on a broad set of target clients, rather than on pure innovation. As the markets for rack-optimized servers and blade servers gradually converge, and new market opportunities such as multinode servers emerge, mainstream server vendors with a strong natural ability to execute will increasingly target the modular server market.
Visionary vendors in this market will either represent the discontinuous leading edge of the market or be large vendors with a plan to drive market success through technology innovation and a narrower product portfolio.
This is a market that addresses specialized edge niches of the broader server market well. This will naturally drive innovation by vendors that may be gaining market momentum, but that only address certain geographies, vertical markets (such as HPC, analytics or cloud infrastructure) or specific workload situations. Consequently, this Magic Quadrant will always have a strong complement of Niche Players that drive real innovation in their target markets, but whose small size or narrower geographic concentration forces them to focus their energies to maintain relevance and deliver business value.
Blades have always represented a transitional stage in the evolution of servers as separate, discrete designs (especially towers and frames) give way to modular designs, and are usually — but not always — the foundation of FBI strategies that further blur the boundaries among servers, storage and networking. This creates an increasing overlap of functionality among product categories that were previously more clearly delineated. In reality, all blade servers are simple examples of a fabric-enabled architecture, due to the switch-based backplane or midplane they exploit. Multinode servers, while chassis-based like blades, are usually based on the same rules for I/O connectivity as those of rack-optimized servers. Blades and multinode servers represent a more proprietary investment than rack-optimized servers, due to the lack of hardware form factor interoperability standards, and (in the case of blades only) the growing dependence on proprietary management tools and virtual I/O solutions.
Due to their modular nature, blade servers offer compelling facilities-oriented benefits, such as improved cabling, rapid hardware provisioning (including the ability to replace failed components), high computing density, energy-efficient design and increasing management automation. However, blades deliver few, if any, incremental application benefits, compared with their rack-based peers. Blades are also not the only choice for modular deployment; rack-optimized servers deliver some modularity benefits, and multinode servers now represent even higher-density deployment for workloads like analytics that demand extreme levels of horizontal scaling. Because some vendors position their blade and multinode servers as part of a standard modular server portfolio, the nascent market status of multinode servers continues to be reflected in this year's blade server Magic Quadrant (even though multinode servers are not strictly blade servers in our opinion).
Blade servers (and some multinode servers) represent a much greater lock-in effect than regular rack servers impose, and ROI calculations need to be more stringently applied. When calculating TCO, users need to consider the longevity of the blade chassis as the most significant investment. Bought at the right time in its life span, a chassis should provide several years of active life and the ability to incrementally add resources or refresh obsolete components over time. Users should carefully match their blade needs and investment objectives to vendor portfolios, product life cycles and vendor strategies for modular architectures as a whole.
The overall server market is gradually transitioning toward a modular data center infrastructure — from building to rack to individual compute elements — that will mask (and ultimately hide) the barriers between discrete compute, storage and networking technology classes, and the management thereof. Blades are not an essential part of this technology convergence (as Oracle is demonstrating with its own integrated system strategy based on rack-optimized technology), but the modular nature of blades makes them a natural fit for the trend, and the blade market has grown steadily as a result. That growth faltered in 2011 and 2012, due to the emergence of multinode servers, although vendor FBI strategies are expected to drive renewed market growth between 2013 and 2015. Current market interest is focused on new-generation workloads such as extreme analytics and in-memory databases, where blades are often overengineered for the task. This low-end cannibalization of the blade server market will continue, and will force vendors to position blades higher up on the food chain, resulting in greater conflict between the optimum use of blades and rack-optimized designs.
Ability to Execute
Product/Service: Core goods and services offered by the vendor that compete in/serve the defined market. This includes current product/service capabilities, quality, feature sets, skills and so on, whether offered natively or through OEM agreements/partnerships as defined in the market definition and detailed in the subcriteria.
Overall Viability (Business Unit, Financial, Strategy, Organization): Viability includes an assessment of the overall organization's financial health, the financial and practical success of the business unit, and the likelihood that the individual business unit will continue investing in the product, will continue offering the product and will advance the state of the art within the organization's portfolio of products.
Sales Execution/Pricing: The vendor's capabilities in all presales activities and the structure that supports them. This includes deal management, pricing and negotiation, presales support, and the overall effectiveness of the sales channel.
Market Responsiveness and Track Record: Ability to respond, change direction, be flexible and achieve competitive success as opportunities develop, competitors act, customer needs evolve and market dynamics change. This criterion also considers the vendor's history of responsiveness.
Marketing Execution: The clarity, quality, creativity and efficacy of programs designed to deliver the organization's message to influence the market, promote the brand and business, increase awareness of the products, and establish a positive identification with the product/brand and organization in the minds of buyers. This "mind share" can be driven by a combination of publicity, promotional initiatives, thought leadership, word-of-mouth and sales activities.
Customer Experience: Relationships, products and services/programs that enable clients to be successful with the products evaluated. Specifically, this includes the ways customers receive technical support or account support. This can also include ancillary tools, customer support programs (and the quality thereof), availability of user groups, service-level agreements and so on.
Operations: The ability of the organization to meet its goals and commitments. Factors include the quality of the organizational structure, including skills, experiences, programs, systems and other vehicles that enable the organization to operate effectively and efficiently on an ongoing basis.
Completeness of Vision
Market Understanding: Ability of the vendor to understand buyers' wants and needs and to translate those into products and services. Vendors that show the highest degree of vision listen and understand buyers' wants and needs, and can shape or enhance those with their added vision.
Marketing Strategy: A clear, differentiated set of messages consistently communicated throughout the organization and externalized through the website, advertising, customer programs and positioning statements.
Sales Strategy: The strategy for selling products that uses the appropriate network of direct and indirect sales, marketing, service, and communication affiliates that extend the scope and depth of market reach, skills, expertise, technologies, services and the customer base.
Offering (Product) Strategy: The vendor's approach to product development and delivery that emphasizes differentiation, functionality, methodology and feature sets as they map to current and future requirements.
Business Model: The soundness and logic of the vendor's underlying business proposition.
Vertical/Industry Strategy: The vendor's strategy to direct resources, skills and offerings to meet the specific needs of individual market segments, including vertical markets.
Innovation: Direct, related, complementary and synergistic layouts of resources, expertise or capital for investment, consolidation, defensive or pre-emptive purposes.
Geographic Strategy: The vendor's strategy to direct resources, skills and offerings to meet the specific needs of geographies outside the "home" or native geography, either directly or through partners, channels and subsidiaries as appropriate for that geography and market.