High Performance Computing (HPC) may be provided via a supercomputer or via parallel processing techniques such as leveraging clusters of computers to aggregate computing power. HPC is well-suited for applications that require high performance data computation such as certain financial services, simulations, and various R&D initiatives.
The market is currently dominated on the demand side by large corporations, universities, and government institutions by way of capabilities that are often used to solve very specific problems for large institutions. Examples include financial services organizations, government R&D facilities, universities research, etc.
However, the Cloud-computing based as a Service model allows HPC market offerings to be extended via HPC-as-a-Service (HPCaaS) to a much wider range of industry verticals and companies, thereby providing computational services to solve a much broader array of problems. Industry use cases are increasingly emerging that benefit from HPC-level computing, many of which benefit from split processing between localized device/platform and HPCaaS.
In fact, HPCaaS is poised to become much more commonly available, partially due to new on-demand supercomputer service offerings, and in part as a result of emerging AI-based tools for engineers. Accordingly, up to 45% of revenue will be directly attributable to the cloud-based business model via HPCaaS, which makes High-Performance Computing solutions available to a much wider range of industry verticals and companies, thereby providing computational services to solve a much broader array of problems.
In a recent study, the Research and Markets team conducted interviews with major players in the market as well as smaller, lesser known companies that are believed to be influential in terms of innovative solutions that are likely to drive adoption and usage of both cluster-based HPC and supercomputing.
In an effort to identify growth opportunities for the HPC market, Research and Markets investigated market gaps including unserved and underserved markets and submarkets. The research and advisory firm uncovered a market situation in which HPC currently suffers from an accessibility problem as well as inefficiencies and supercomputer skill gaps.
Stated differently, the market for HPC as a Service - e.g. access to high-performance computing services - currently suffers from problems related to the utilization, scheduling, and set-up time to run jobs on a supercomputer. The company identified start-ups and small companies working to solve these problems.
One of the challenge areas identified is low utilization but (ironically) also high wait times for most supercomputers. Scheduling can be a challenge in terms of workload time estimation. About 20% of jobs are computationally heavy 30% of jobs cannot be defined very well in terms of how long jobs will take - within 3-minute window at best. In many instances, users request substantive resources and don't actually use computing time.
In addition to the scheduling challenge, we also identified a company focused on solving additional problems such as computational planning and engineering. Research and Markets spoke with the principal of a little-known company called Microsurgeonbot Inc. - doing business as MSB.ai, which is developing a tool for setting up computing jobs for supercomputers.
The company is working to solve major obstacles in accessibility and usability for HPC resources. The company focuses on solving a very important problem in HPC: Supercomputer job set-up and skills gap. Their solution known as "Guru" is poised to make supercomputing much more accessible, especially to engineers in small to medium-sized businesses that do not have the same resources or expertise as large corporate entities.
For more information about this report you can visit the Research and Markets website.