Microsoft tackling cluster computing
Microsoft has entered the cluster computing market with hopes of grabbing a share of the market currently led by Linux. The proliferation of clusters for heavy duty computing continues across many business segments, from industry to government, as evidenced by the latest edition of the Top500, which shows that cluster systems comprise 60% of the list.
According to this article from Grid Computing Planet, Microsoft’s initial software entry works on clusters of up to 128 machines and Microsoft intends to better integrate heterogeneous applications on the cluster compared to Linux as well as offering better support. An overview of their cluster solution can be found here.
The new software will include an open source MPI middleware. That’s right - open source. More info about Microsoft’s decision to implement this into the Compute Cluster Solution can be found in this eWeek piece.
September 20th, 2005 at 11:49 pm
I talked to one of the people inside the HPC team at Microsoft last week, and they are hoping to win back some users with the rollout. If done right, the MS HPC product could indeed be very powerful and well received in the community. There are many small departments that have applications that could use a little more computational power, and since nearly every infrastructure already has XP in place, adding some HPC-XP servers would be easy.
However… the problem for MS is that the genie is already out of the bottle. Linux is a serious world-wide competitor, and most supercomputers on the planet run Linux. To take back market share, MS needs more than a niche product aimed at companies insufficiently tech savvy to buy a Linux cluster from IBM.
No, the product has do something *better*, in addition to adding security concerns, administration headaches, and configuration problems. New products that have staying power have an edge, a unique trait that puts them way out ahead of the competing solutions. Being an “also ran” does not win back market share — it can slow errosion, but not win it back.
So the question for HPC-XP from MS remains…. what does it do *better* (and cheaper)? White papers from paid “industry analysis” on cost of ownership need not be cited
Supercomputer center directors and NSF program mangers often seek the perfect sound bite “Without the gigablaster2000 supercomputer at your center, I could never have mapped the genome of the blue-eyed snipe”. So, what will be Bill’s sound bite at SC05 regarding his new product rollout, and will he practice looking as cool as Jobs pulling an ipod nano from his change pocket?