Several years ago, in 2019, Hewlett Packard Enterprise made a public vow to offer all its offerings “as-a-Service” by 2022. HPE’s GreenLake cloud services portfolio is essential to this strategy by HPE offering organizations a highly flexible, scalable, secure and fully managed and monitored IT experience. I’ve been following the company’s strategy closely and have written on it on multiple occasions (see here and here). Continuing its migration of offerings to everything-as-a-service, last week, HPE announced that its high-performance computing (HPC) offerings will now be available through GreenLake—a move designed to make it easier for mainstream enterprises to adopt HPC. Let’s take a closer look.
Lowering the bar for entry
Advancements in high-performance computing have opened up a whole world of possibilities in recent years for researchers, engineers, enterprises and more. With the massive compute power of these systems, organizations can develop models, run simulations and derive insights from their reservoirs of operational data—most of which were previously too large and complex to tap. These systems are also ideal for next-generation technologies such as AI and automation, which are becoming more and more common. HPC’s power can aid in everything from the research of vaccines, to the development of household consumer products, to designing safer vehicles. These systems can run simulations that predict stock trends, accelerate and improve the safety of oil and gas explorations and more. With all this potential, it’s no wonder the HPC market is estimated to grow over 40% by 2024 (according to Intersect360 Research).
One issue, though, that must be dealt with before HPC reaches widespread, mainstream usage in the enterprise is that these robust computer systems are costly, complicated and resource-draining. The average business likely does not have the time, money or personnel to deploy and manage these HPC systems.
This is where HPE GreenLake stands to revolutionize and accelerate the deployment of HPC in the mainstream enterprise sector—by delivering HPC via a pay-as-you-go, scalable, fully-managed service. HPE says this will hasten the deployment of HPC projects by as much as 75% and reduce costs by as much as 40%. These improvements put this game-changing technology way more in reach for the average business.
To start, HPE GreenLake will offer an HPC service based on the company’s Apollo systems, followed by the rest of HPE’s strong HPC portfolio (including Cray systems) in the unspecified future. This first offering leverages networking and storage technologies and is built explicitly for modeling and simulation workloads. Additionally, it features HPC workload management software, HPC cluster management and monitoring and support for HPC-specific containers and orchestration.
Customers will also gain access to some of GreenLake’s other perks. The HPC service includes HPE GreenLake Central, a platform for the management and optimization of customers’ HPC services; Consumption Analytics, which provides customers a real-time view of their GreenLake usage and costs; a self-service dashboard with which customers can deploy and manage their HPC clusters with a simple point-and-click, without causing any disruption to the workflow; and lastly, HPC, AI and App services, which serve to standardize and package HPC workloads inside of containers. That last feature promises to simplify the modernization, transfer and access of data—a massive part of any company’s digital transformation.
The new service will be available to the general public this coming spring in three different size configurations—small, medium and large. Upon ordering, HPE says organizations will receive their bundles as soon as 14 days, with their fully-managed HPC service up and running shortly after that.
Broad ecosystem support
Crucial to any new technology, HPE’s GreenLake Cloud Services for HPC benefits from a supporting cast of partners. Those who lack the physical real estate to house the HPC systems and equipment have the option of deploying them in one of HPE’s colocation partners’ energy-efficient, scalable data centers. The customer can then access its HPC services remotely through GreenLake. This option is for enterprises who don’t want datacenters or the hassle of dealing with the physical infrastructure of a datacenter.
Additionally, HPE has a lineup of ISV partners (independent software vendors), including Activeeon, Core Scientific and TheUberCloud, who provide software for a wide range of potential HPC applications—AI, automation, analytics, computer-assisted design (CAD) and more.
I imagine this announcement is welcome news for any businesses that are curious to see what doors HPC will open but wary of the significant overhead. Organizations of all sizes will be able to reap HPC’s benefits without the headaches traditionally associated with such a deployment. There’s a reason these as-a-Service models have taken off in recent years—if an organization outsources the wrangling of IT to a third party, it frees up resources and personnel to focus on more strategic business goals. HPC, in particular, seems like a no-brainer for outsourcing, given its costs and complexity. All in all, I believe HPE GreenLake Cloud Services for High-Performance Computing represents the welcome democratization of HPC for the average enterprise. It should do much to accelerate the widespread enterprise adoption of HPC. I think this is a good vertical for HPE to focus upon given its pedigree in high performance infrastructure and its investment in data management tools specifically for HPC.
Note: Moor Insights & Strategy writers and editors may have contributed to this article.