High-Performance Computing with the Cloud that Comes to You: HPE GreenLake for HPC as a Service

[ServicesExperts]

You don’t have to be a massive research organization to take advantage of high-performance computing. You can choose a pay-per-use solution right-sized for your business, without the need to make major upfront investments.

By Mark Seamans, Senior Director, Platforms & Horizontal Workloads aaS, HPE GreenLake Cloud Services

One of the truly exciting things about high-performance computing is the sheer range of use cases around some of the world’s biggest challenges. HPE has customers who are using HPC for projects ranging from long-term climate simulation to help predict global warming, to wind tunnel simulations to design more efficient blades for wind-based power generation, to genomics and healthcare breakthroughs.

More and more enterprises are benefiting from the pioneering work of the most advanced scientific and engineering applications of HPC in high-profile projects like these. Companies are starting to realize that you don’t have to be a massive research organization to take advantage of HPC capabilities to advance whatever you’re working on. Whether it’s for computer-assisted engineering in automobile design, fraud detection in the financial sector, or personalized healthcare, HPC is available at a scale that's much more accessible today than ever before.

The rise of the pubic cloud has certainly played a part in that. Many companies use cloud for mature workloads that have been distilled down to a high degree of repeatability. However, HPC, in general, has traditionally been associated with data center and on-premises environments, and with initiatives that involve cutting-edge technologies. Low-latency networking is a good example. HPE recently brought to market HPE Slingshot Interconnect, a very low-latency, super-high-speed networking environment. Highly demanding workloads like this tend to run more naturally and are more available in on-premises or co-location environments than in the public cloud.

There’s an economic aspect, too. The nature of high-performance computing is that you're typically running workloads flat out – you want to exploit every ounce of compute and network capability, and in some cases, you’re really driving hard on storage. If you’re doing that 24x7 in the public cloud, you can end up with a pretty significant bill.

Using on-premises equipment can be a more cost-effective way to do it. What’s more, now you can go the on-prem route without the need for big upfront investments with HPE GreenLake, our self-service, highly scalable, pay-per-use platform.

An as-a-service approach to on-prem HPC

Here’s how it works. You choose from “t-shirt-sized” configurations – small, medium or large – to meet your workload needs. The order is processed, and our supply chain sends out the ready-to-go solution in as little as 14 days. Our experts install and configure it on-site and deploy any necessary apps. You cut out procurement and deployment delays and get up and running quickly. It’s a true pay-per-use solution; we apply HPE metering technologies to determine your usage of the infrastructure, and we bill you monthly based on that usage.

HPE works behind the scenes with management services and automation to monitor and operate the solution to keep it running at its peak, and we take care of versioning to keep the infrastructure current. In addition, HPE GreenLake Central, our self-service software platform, delivers the information and insights you need to govern your usage. (Read how HPE GreenLake for HPC delivers accelerated outcomes and financial value on your terms.)

5 big advantages of an on-premises, cloud services approach

HPE GreenLake for High Performance Computing gives you:

1. Access to the right level of expertise. It’s important to consider HPE GreenLake for HPC not just from a technology perspective, but also from a services perspective. It’s fully integrated software running on a curated hardware platform, and it comes with built-in operational support. If there’s a component that needs to be replaced, for example, we do that for you.

If you choose, we can also provide a higher level of solution management and optimization through HPE GreenLake Management Services (GMS). For example, if you’re just getting into HPC for specific workloads, we can provide assistance to help you through the process of setting up your job runs, executing your jobs, optimizing, and troubleshooting. With GreenLake, accessing this higher level of service capability is not an extra charge – but rather can be incorporated into the consumption costs for the HPC assets that you are consuming.

Access to the right expertise can make all the difference for an HPC implementation. We find that many customers who are considering HPC-style workloads may not have done anything similar in the past, and the expertise and experience that HPE GMS can provide can deliver significant improvements in the ROI that customers obtain from their HPC investment.

2. Workload-specific solutions. In addition to providing a strong foundation for running HPC in general, HPE also addresses specific workloads with solutions for some of the common things that organizations do with high-performance computing. For example, with our initial release, we targeted manufacturing and computer-assisted engineering (CAE). Together with our partners, we can help you support workloads for things like computational fluid dynamics (CFD) for wind tunnel simulation; structural mechanics for designing a bridge or a building; and computational electromagnetics for designing, say, a cell phone antenna.

3. Flexibility of location. You may decide that you don’t want to add HPC equipment to your own data center’s physical footprint. No problem; we can install the solution in a co-location facility.  As with other aspects of GreenLake, the costs for the co-location deployment can be built-in to the consumption charges for your solution – so that you are able to “start small” and scale as the size and capabilities of your solution grow over time.

4. Consumption analytics and recommendations. I mentioned HPE GreenLake Central, and we’re finding that this is an especially attractive feature for many companies, whether they have long experience in HPC or they’re venturing into it for the first time. It’s a one-stop shop for customers to access the system and do all kinds of things relating to the measurement of the economic value of their solution. Powerful capabilities for consumption analytics are built in. You can go in and view your usage and billing for the system. You can drill into details on what's driving your consumption. The system will even make recommendations – “here are things you could be doing that would drive down your cost of consumption” – to help you maximize the results per dollar of investment that you make into the solution.

5. Application portability with containers. With HPE GreenLake for HPC, you can run your applications on containers, which enables portability from other environments and positions your workloads for this new style of application deployment and management.

Unlocking what’s next for HPC

With our long history in high-performance computing and the combination of HPE with Cray, HPE is an extremely strong and experienced player in the HPC market. We can help you achieve your goals whatever your level of experience, and whether you’re just getting into HPC initiatives or you’ve been doing HPC for decades.

We’ll continue to drive innovation in this space, such as leveraging the low-latency networking capabilities of the Slingshot technology that I mentioned earlier. We also see tremendous customer interest in blending HPC with AI and machine learning as organizations take advantage of those powerful computational techniques to deliver game-changing business outcomes.

Join us for this HPE World Watch webinar on May 12, 9am PT, to learn how to combine an agile, elastic, pay-per-use cloud experience with the power of market-leading HPC systems: Leverage the Power of HPC Delivered as a Service. Learn how you can speed up deployment of HPC projects by up to 75% and reduce capital expenditures by up to 40%, while harnessing your data faster and accelerating time-to-insight.

Learn more about HPE GreenLake high performance compute

Watch a short Chalk Talk video about HPE GreenLake for HPC – a great overview of the features and benefits.

Related article:  HPC as a service: high-performance computing when you need it via HPE's Enterprise.nxt

Stay current on top tech trends and expert advice. Sign up for weekly insights and resources

Mark Seamans is Senior Director, Platforms & Horizontal Workloads aaS, for HPE GreenLake Cloud Services.

Services Experts
Hewlett Packard Enterprise

twitter.com/HPE_Pointnext
linkedin.com/showcase/hpe-pointnext-services/
hpe.com/pointnext