HPE this morning announced an HPC-as-a-service offering that sits at the confluence of two well-established market trends: broadening adoption of HPC in the enterprise, and accelerating demand for cloud-based HPC.
The pay-per-use HPE GreenLake cloud services for HPC runs on-premises or in a colocation facility and is designed to avoid the complexity and cost of traditional HPC deployments “by delivering fully managed, pre-bundled services based on purpose-built HPC systems, software, storage and networking solutions,” HPE said. The services are ordered through a self-service portal with point-and-click functions to choose workload configurations, which the company said can be received “in as little as 14 days.”
The HPC service will initially be based on HPE Apollo systems combined with storage and networking technologies built for running modeling and simulation workloads. Bundles are offered in small, medium or large configurations.
The offering includes:
- GreenLake Central, a software platform for customers to manage their HPC service
- Self-service dashboard to run and manage HPC clusters through a point-and-click function
- Consumption Analytics to monitor usage and cost based on metering through HPE GreenLake
- HPC, AI & App Services to packages HPC workloads into containers, designed to ease transfer and access to data
The announcement comes as part of HPE’s overall strategy, announced last June to transition into an as-a-service company for its entire portfolio through subscription-based, pay-per-use offerings by 2022. The company also will provide its products in a capital expenditure- and license-based model.
“The massive growth in data, along with artificial intelligence and high performance analytics, is driving an increased need for HPC in enterprises of all sizes, from the Fortune 500 to startups,” said Peter Ungaro, HPE senior vice president and general manager, HPC and Mission Critical Solutions. “We are transforming the market by delivering industry-leading HPC solutions in simplified, pre-configured services that control costs and improve governance, scalability and agility through HPE GreenLake. These HPC cloud services enable any enterprise to access the most powerful HPC and AI capabilities and unlock greater insights that will power their ability to advance critical research and achieve bold customer outcomes.”
HPE, which last year completed its acquisition of supercomputing stalwart Cray, is ranked no. 1 in HPC hardware; it also is the leading systems provider in the American effort to stand up exascale-class supercomputers starting next year. The overall HPC industry is in growth mode, projected to expand by more than 40 percent to almost $55 billion in by 2024, according to industry analyst firm Intersect360 Research. But as HPE pointed in its announcement, traditional deployment and management of HPC systems is costly, complex and resource-intensive. Top concerns involve system costs, operational costs related to power and cooling, and lack of skilled HPC technical staff. HPE said analyst firm Forrester Research found in a commissioned study that GreenLake as-a-service cloud platform speeds up HPC project deployments by up 75 percent and reduces capital expenditures by up to 40 percent.
The company cited the customer testimonial of Zenseact, a software developer for autonomous driving solutions based in Sweden and China, which analyzes its network of test vehicles and software development centers.
“By deploying HPE’s high performance computing solutions as-a-service with HPE GreenLake,” said Robert Tapper, Zenseact CIO, “we are addressing our mission by performing 10,000 simulations per second, based on driving data from our test cars, to accelerate insights for designing software to enable safe autonomous vehicles.”