The IDC has just released a report covering the HPC server market growth in the third quarter of 2007. According to the report, factory revenue grew 8.8% from Q2 of this year. This is up 18% from Q3 of 2006. The worldwide HPC server market grew 0.5% to reach $13.1 billion in Q3. The report cites a lower initial entry cost for small businesses and research organizations to be the key driver in the growth curve.
The rapid growth of HPC server revenue and processor counts since 2002 fits the classic profile of a disruptive technology,” said Vernon Turner, senior vice president of IDC’s Enterprise Infrastructure, Consumer, and Telecom research.
A few interesting figures from the report include:
.: Over the last four years the HPC market has seen revenue growth averaging 20% per year.
.: HPC clusters continued to gain momentum across all HPC segments. Revenue from clusters represented 68% of the overall HPC server revenue for the third quarter of 2007.
.: The fastest-growing area is the workgroup segment for systems priced under $50,000, which is projected to have 11.4% CAGR through 2011.
.: In 2006, HPC server revenue surpassed $10 billion. IDC values the broader HPC market, including servers, storage and services, at $16.3 billion in 2006.
What does this tell us? We all see the NSF Track 2 scale system procurements in the news, yet the >$50,000 is dominating the market growth. Simply put, HPC is becoming more accessible to the “average-joe” user. Until recently, procuring and operating an HPC system of any magnitude was not a trivial task [and arguably still isn't]. Beowulf clusters running Linux certainly helped close the gap between the big-iron and small engineering/scientific firms. However, recent advancements in HPC [specifically cluster] system management from the likes of Rocks, LNXI and Appro have allowed the small pockets of users to focus on the science surrounding the development and execution of their applications, rather than system deployment/management. As systems software packages continue to grow and improve, I expect we’ll see a further parallel growth in small HPC deployments.
Read the full report here.