The Democratization of HPC
The earliest supercomputers took the form of the massive, expensive machines that came out after World War II. The size and cost along with the complexity of operating proprietary systems limited the user pool to large academic, government and industrial research facilities, leaving most organizations without access to the tremendous benefits of HPC.
When x86 servers came into their own in the late 1980s, they enabled the creation of HPC clusters that greatly improved the speed and flexibility of these systems and became the dominant style of HPC. Today, modern technologies, standards and open source software are starting to change all that, enabling a whole new group of mainstream HPC users and revolutionizing the way companies use data and develop products.
As the commoditization of IT expanded over the years, it has had a tremendous impact on HPC. Companies used to build everything from the ground up for their supercomputers, which created complex, proprietary and expensive systems. Now, most of the underlying technologies are standardized from Intel and other component providers, making it easier to create HPC solutions fast and affordably.
As the commoditization of IT expanded over the years, it has had a tremendous impact on HPC
Technology providers also are able to design systems for various applications and particular jobs. The OpenHPC community is creating standards for how HPC software is developed, and open source options are now available to enable vendors and customers to easily build stacks. This software standardization has eliminated a major barrier to adopting HPC, allowing companies with limited IT resources to take full advantage of the technology.
The HPC market is broadening as non-traditional and smaller companies are coming around to the idea that HPC can have a major impact on innovation and competitiveness. We’re starting to see a transition from HPC as a monolithic, intimidating technology to a valuable tool for specific business processes and industries, such as consumer packaged goods, aircraft design, manufacturing, sports and life sciences.
Modern day HPC providers are expanding the market by eliminating difficult, expensive, complicated facets and delivering solutions that let customers focus on specific goals. Even with its significant progress and increased accessibility, many CIOs still don’t recognize HPC not only as a realistic option, but likely the best one available.
Barriers to mainstream adoption have remained, not the least of which is the ongoing perception that HPC is too complicated and focused only on bleeding edge, sophisticated scientific research. But the evidence clearly shows that the future of business productivity will lean heavily on HPC, which enables the testing of products virtually instead of physically.
According to the National Center for Manufacturing Sciences, 98 percent of all products will be designed digitally by 2020, yet 94 percent of the 300,000 small and medium-sized manufacturing companies in the U.S. have little or no HPC expertise. However, some innovative companies in various industries are starting to use HPC for a host of non-traditional processes, including packaging design, assembly line efficiency, weather forecasting, genome sequencing and more.
For example, some of our academic customers have developed programs for the small and medium-sized manufacturing companies in their areas to educate them on the benefits of HPC and provide access to HPC systems to get them started. This is happening with the purpose of increasing manufacturing innovation and competitiveness.
As CIOs embrace HPC, one big caution or trap to avoid is getting enamored or overwhelmed with the technology. It’s rarely necessary to find the fastest, most powerful system as opposed to one that addresses your company’s specific business issues or outcomes you are trying to drive.
When evaluating HPC systems, simply ask yourself, “what strategic objectives need to be met, and what three things can help with innovation and competitiveness to get there?” Then find an industry-specific, purpose-built solution that shields complexity and addresses your business and outcome needs.
Cloud HPC also can be an effective option for some CIOs, but it’s not for everyone. For companies that only need HPC resources occasionally, cloud can be an affordable and effective solution. But if you’re using HPC constantly to analyze data, cloud can actually be more expensive than an on-premises data center. The hybrid cloud approach is best for most companies, providing flexibility for additional resources as needed and guaranteed use of your own systems for mission-critical operations. It’s important to have a dialog with a trusted vendor advisor on the best set of HPC services to fit your unique set of requirements.
The bottom line for CIOs is that HPC is no longer an expensive, fringe technology reserved for the most elite scientific research organizations. Its time has come for mainstream business use, and those companies who choose not to embrace it do so at their own peril as competitors quickly jump on the bandwagon and improve their competitive position.
By carefully selecting HPC solutions that address specific business needs and meet their company’s budget and logistical needs, CIOs can use HPC to take their operational efficiency to the next level, and jumpstart innovation to position their companies for ongoing success in the marketplace.