Customers still have a crush on IBM and Hewlett-Packard as x86 server vendors while Sun Microsystems and Dell continue to receive lukewarm satisfaction, according to a new survey. The latest "Vendor Faceoff" study of enterprise x86 customers by Gabriel Consulting Group (GCG) shows IBM holding a narrow lead over HP in favored …
Web Poll, not Survey
In the past, Gabriel Consulting did anonymous web polls to gather data, not scientific surveys. GC offered free web coupons to those who completed their poll. In the past, GC advertised the poll on second tier IT news web sites. The only mechanism they used to prevent "stuffing the ballot box" was a unique email.
Regardless of GC's claims of the power of its statistical analysis methods, no statistical analysis can make unscientific data scientific.
While it is certainly possible GC has changed its data collection method to a scientific one, there is no way to know this, as no reporter has asked Daniel Olds how he collects his data.
This is Hardly a Scientific Survey
Surveying this few customers gives us an insight into what's going on in large enterprises, but it is hardly a scientific survey and should not be interpreted as such. Gabriel consulting? Who are they? Let's stick with the big guns in this space that customer really listen to, such as Gartner, IDC and TBR.
Arguments aside, let's take a look at the four vendors mentioned:
Sun - Sun has been in and out of the X86 business over the last few years and despite a much balleyhood reentry with AMD chips, their market share has barely increased. The latest Sun servers appear to look good, but they are very expensive in comparison to their peers. Also, more compact, but can a customer power and cool a full rack of servers of this format?
IBM - IBM's market share of X86 space continues to be squeezed by HP and Dell and they continue to be a distant third. IBM is aservices-led company which - going all the way back to the Gerstner days - sells value in the 'solution' rather than the hardware itself. For example, IBM blades look very dated in areas of energy efficiency and high availability features when comparing to HP, Dell and Sun's current offerings.
HP - HP is #1 in X86 unit share worldwide, holding market share doing well in this space and their new blade product is very credible, albeit an expensive option. HP's done a good job of wrapping services around their products and continues to lubricate well the reseller channel that has brought them so much of this business.
Dell is #2 in X86 in units and is increasing market share faster than the other three vendors. Their product line is well established and they have #1 market share in the US. Dell has innovated significantly around virtualization, power and cooling, won the recent InfoWorld "battle of the blades" and continues to be first to market to ship new Intel/AMD technology due to their business model. Dell has just embarked upon building up their channel/reseller model.
Conclusion: Customers speak with their wallets. I would guess that IBM and HP's prowess in this unscientific survey comes off the back of things that these companies are doing elsewhere in the customer's environment, such as Unix, mainframe, services and so on, so that their overall business perception is playing into the X86 perception. Same with Sun - there's still a lot of old big iron Solaris out there.
Dell's the only pure play - that says a lot about this company's focus and where it is leveraging its R&D dollars. Sounds like a lot of these data center managers should shake things up and take a look at competitor products.
HP is the company to beat
The anonymous poster above (#2 comment), you have made a number of claims without providing much of convincing evidence, if any. I think there are a number of errors in the way you described each vendor. Allow me to refute your claims:
The King of x86 server space. HP has great momentum and I think anytime any customer makes a purchasing decision, there is no way they can ignore HP without giving a very hard look -
HP has a long x86 server history, a huge install base (especially when you combine with Compaq), they know the x86 space in and out, have the right mix of R&D to solve customer problems efficiently that a DELL can never do for example and are able to provide their products at a very attractive price, in many cases even beating DELL - just to name a few. They invest in R&D at all levels beginning with hardware, firmware, software and the momentum they have in the marketplace shows that the return is huge. HP has a great presence in the volume consumer space similar to DELL and they are able to offer lower price compared to others and even matching or beating DELL in a few cases.
The primary weakness for IBM is that their products standalone are more expensive compared to others, even compared to Sun in some cases. IBM can match or exceed HP in many scenarios in terms of being able to offer the complete solution, and they can go one step ahead in terms of being able to take over the entire IT operations of many big companies.
Over the last few years DELL has discovered that low price isn't everything a customer looks for. Their absence from AMD server space have hurt them a lot, specially against HP. While it is true that DELL has grown faster compared to others in the last quarter (well not all, actually Sun has grown percentage revenue faster in x86 servers but their market share is so small that it probably isn't even worth mentioning), their products are still weak compared to HP. Even now DELL competes primarily on price. HP is very focused on Blade and have a better solution compared to others, just look at how fast their blade market share has grown after releasing C-class blades last year, while DELL has trailed further and further in the blade market. DELL has a dubious track record on blade, they claim that blade isn't good for everything (well, may be for valid reasons) and this is exploited by HP and IBM. Notwithstanding what Infoworld's product review says, DELL is way behind HP and IBM in terms of offering a complete blade package - note the blade market share speaks volume about this claim. BTW, the way you are beating the DELL drum, looks like you either work for DELL or are a huge DELL fan - (no offense intended). They are trying to build a channel model, but this is not easy and often in conflict with their direct sales model. Well, we will see if DELL can continue to do as well in x86 server space as they did last quarter, I am certainly hopeful.
Your claims about Sun are barely true. While their market share didn't increase much in terms of overall x86 server market share, Sun's x86 server revenue actually increased almost 50% in the last quarter. They have gone to 5th position in terms of x86 server revenue and right behind Fujitsu-Siemens. I think they did pretty good, in spite of having a number of weaknesses compared to others:
- Sun doesn't sell Windows server yet on their hardware, so they already were not even reaching more than 50% of the market.
- Sun didn't have any Intel product line, this already made their offerings uncompetitive compared to others, specially with quad-core Intel CPUs selling very hot over the last few quarters.
- Sun didn't have a general purpose blade offering until last quarter.
- Weak storage offerings
- Expensive compared to HP and DELL, but less expensive compared to IBM.
- Sun's servers are also top rated in Inforworld reviews, much like how DELL winning the so called "battle of the blades" - but CIOs have to look at many other aspects when making a purchasing decision. The Infoworld review does not even look at the compelling features HP blades provide compared to others.
While some of these issues are gradually being fixed by Sun, they will remain a distant player in x86 server market for the foreseeable future, but they will continue to grow their x86 business at a healthy rate.
Bottom line is, I would generally agree with the findings of the survey. HP and IBM are perceived much more credible server vendors in terms of being able to offer the complete top-to-bottom infrastructure and do them with competence. DELL will still need to continue working on their credibility problems. HP is very well positioned in the x86 server market, but they need to continue innovating to be in the limelight. IBM will keep doing well based on their ability of offering everything, and Sun has to struggle very very hard to gain more x86 credibility.
blades aren't everything
anonymous poster #3 - although currently the sexy and cool trend in the X86 space, blades make up <10% market share for X86 servers and are not a good fit for many environments. Most data centers cannot leverage the density due to power limitations to each rack, or being unable to deliver sufficient cooling. HP has some very credible product sets but cannot lie by blades alone. The article today on this website http://www.theregister.com/2007/10/04/gartner_2011_significant_disruptions/ indicates why the average customer cannot even half-populate their racks with rack-dense servers, let alone blades.
Blades are the fastest growing segment of the x86 market
It may be true data center cannot take advantage of the density allowed by blades. However, they will go for all the density they can, and blades offer more compute power per square inch than rack servers.
Also, sharing of components like power supplies and fans allow blades to require LESS overall power and generate LESS overall heat than traditional servers.
Combine that with power management such as HP offers, where procs can be dynamically scaled back based on application requirement (or lack) in conjunction with pre-set limits per rack, it is clear blades (because of intelligent enclosures and software/firmware) have advantages over dumb servers in dumb racks.
I don't see any Tier-1 vendor going with Blades only, but that does not mean they are selling because they are cool or sexy. IT heads are a little too smart (and way too boring) to make decisions based on sexiness.
Re: blades aren't everything
I concur with Jack Pastor regarding blades. While it is true that blades are not an alternative to rack server in many cases, they have advantage from the perspective of overall power consumption, cooling requirements, cable management etc. and in many cases provide more dynamism with respect to workload management. Technical merits aside, the more important point is that blade market is growing at a much much faster rate compared to rack server, so not having a competitive solution means a vendor misses the most growing part of the x86 server market. Note that the study is mostly about vendor perception, so having an elegant blade solution actually makes that vendor a more credible technology supplier. Today the buzzword is blade in the server market, and you can see both HP and IBM are chasing customers very aggressively with blade solutions where a rack solution could have been fine, e.g. the C3000 blade from HP and the recently announced BladeCenter S from IBM are both examples where these companies are trying to provide a package in blade form factor which include pretty much everything an SMB would need. So the blade density advantage has also morphed into lower corners of the market where customers would be looking at these packaged blades even though they may need just a few of them.