Friday afternoon I received and email from HP's PR firm in anticipation of a Sun launch event tomorrow in San Francisco, with CEO Jonathan Schwartz; Hector Ruiz, CEO AMD; John Fowler, Sun EVP of Systems; Andy Bechtolsheim, Sun Systems Designer and some customers.
HP's ambush email is basically trashing Sun's forthcoming new product announcement, which we'll find out about tomorrow. Here is the text of the email:
As you are probably aware, Sun is expected to announce its new Opteron-based system – the Sun Fire x4600 – next Tuesday. Sun is also finally expected to introduce its second attempt at blade servers. What does all of this mean for customers?
Customers want choice for their x86 platforms and Sun can’t deliver
Today 80 times more customers worldwide choose Linux on x86 compared to Solaris on Sun x86 systems. Windows shipments on x86 also outpaces Solaris on Sun x86 by 269 times. (*1)
HP continues to lead the market in x86 shipments across all the major operating systems – Linux, UNIX and Windows (*1)
Customers aren’t asking for scale-up x86
Scalable x86 systems from IBM, Unisys and others have existed for years, but the market data shows that these systems are a very small niche. Here are notable results from the Q106 IDC Quarterly Server Tracker:
- 99.8% (virtually the entire market) of all x86 systems shipped in Q106 were 1 to 4 sockets.
- 8 socket x86 systems accounted for only 0.07% of the x86 units shipped.
- From a revenue perspective, 96% of the x86 server revenue was in the 1-4 socket space, while 8 socket x86 systems accounted for only 1.4% of the market.
One has to wonder what Sun’s moves in scale-up x86 mean for Sun’s UltraSPARC customer base. With no real roadmap for SPARC, will longtime Sun customers be left in the lurch?
Customers who need to scale up can do so with HP Integrity servers that actually provide the balance, the RAS features, the multi-OS support and a leading enterprise UNIX, all in one package.
Is Sun committed to the Blades market – their track-record speaks for itself
Sun announced its entrance into the blade market in early 2003 but by June 2005, their first blade effort had fizzled and they had silently exited the market – leaving the small number of Sun customers high and dry. Heat and I/O problems are rumored to have been at the root of the product’s demise, but the poor performance of the 650MHz USIIi blade and product delays contributed as well.
*1 Source: All of the market share figures are for the 1st Quarter of 2006 and represent worldwide results as reported by the IDC Worldwide Quarterly Server Tracker, May 2006)
I checked in with Sun to get a response. The company doesn't want to preempt tomorrow's announcements, but the following statement came in via email from Graham Lovell, senior direcdtor of Sun's x64 systems product line:
HP should stop confusing customer rejection of Itanium with any lack of demand for scale up systems - we're seeing robust growth, and share gains, across x64 and SPARC platforms.'
And secondarily, of the 5 million Solaris licenses we've distributed in the past year - which is, I believe, more than HP has distributed of HP-UX in its entire history - nearly two thirds were on x64, and a significant percentage of that on HP's x64 servers. Is HP dissuading Solaris customers from considering HP? Dell must be thrilled!
Lovell makes a good point about HP's Itanium bet, and momentum behind Solaris, but doesn't address the Linux on x86 space, blades or the SPARC roadmap. More to come tomorrow when the smoke clears and Sun actually debuts its new products and responds more formally to HP's points.
Here are Gartner's server numbers from Q1 2006. Sun showed growth for the first time in two years, according to Gartner's report.
Maybe HP CEO Mark Hurd is turning up the heat in a war of words with Sun. In March, former Sun CEO (now Chairman) Scott McNealy proposed to a converged HP-UX/Solaris 10 platform to Hurd, citing the end of PA-RISC server development and relegating HP-UX to the faltering Itanium. We haven't seen any progress on that front...