Microsoft has bucked most of the major software makers by agreeing to count multicore processors as a single chip in its product pricing schemes.
Rightly recognizing an opportunity to boast, Microsoft today dedicated an entire press release to its pricing decision. The company will stick with its per processor model for products such as Windows Server and SQL Server regardless of how many cores Intel or AMD manage to pack into a chip. Microsoft's openness on this matter is commendable and in stark contrast to the likes of Oracle and IBM, which tend to shy away from admitting they will price software per core.
Anytime, however, that Microsoft comes off looking rosier than Oracle and IBM (all of whom are equally aggressive) in a pricing issue, you know deeper motives than just customer satisfaction are at hand.
Oracle, IBM, Veritas, BEA and a host of others have considerably more to lose than Microsoft from multicore chips in the short term. For starters, all of these vendors run their software on Unix and Linux servers, which are already running on dual-core chips from IBM, Sun Microsystems and HP. At the most basic level, these software makers would lose half of their revenue on a per processor pricing scheme if they gave in to the chip makers. (IBM is an especially interesting case as it has DB2 and WebSphere code to profit from and makes its own Power5 chip).
None of these enterprise software makers are prepared to sacrifice their wonderful profits just because some over-active chip makers are offering double the performance in the same, old package. This is one reason Sun has moved to a per employee model. Sun doesn't have huge existing software revenue to defend and is looking to tempt customers with a model that makes solid use of multicore chips.
Microsoft, by contrast, is still playing in the single core processor world. AMD isn't expected to deliver a dual-core chip until the later bit of 2005, and Intel won't come to market with a dual-core Xeon until 2006. Intel will have a dual-core Itanium chip next year, but due to Itanium's tepid sales, this shouldn't chuff Microsoft too much.
In addition, Microsoft tends to play on smaller systems. It's not losing out millions and millions because a 64 processor box has suddenly turned into a 32 processor box.
Still, there is no denying that paying for a four-processor SQL Server license and being able to run that on 8 cores is attractive to customers and a potential loss situation for Microsoft. What the company has done, that others aren't prepared to do, is wake up to the realities of the processor market. The chip makers used to boost performance by ratcheting up GHz. Software makers couldn't complain about this. Now, the chip makers are just improving performance by adding more cores to a chip. This trend will only get more and more dramatic as four, eight and 16 cores processors roll out over the next couple of years.
It will also be interesting to see how Red Hat responds to this challenge from Microsoft. A move to Linux could be much less attractive if you get twice the bang for your buck than before on a Microsoft box.
It's doubtful that Oracle, IBM and others will be able to hold out much longer with their old school pricing plans. Microsoft has sided with the chip makers and put some real pressure on the competition. ®