Blog post written by Server Experts blogger Jean-Luc Assor, Worldwide Hybrid HPC/HPC Cloud Program Lead at HPE and posted on 11-11-2018.
What is hybrid HPC? And how does it solve the three main HPC challenges that many enterprises face?
High-performance computing (HPC) is growing at an unprecedented rate. Many corporations that were using on-premises data centers are exploring cloud computing technologies and deployment models. The reasons include wanting to improve the end-user experience, expand data center capabilities and take benefit of consumption-based models that allow for a reduction of upfront investments while helping to speed up deployment.
What's more,, across every industry, businesses are under severe competitive pressures to deliver more custom products and services through digital platform business models and rely on HPC system to boost the capabilities of their Information Systems. In fact, IDC estimates that worldwide cloud revenues will reach $554 billion in 2021, more than double those of 2016 and cloud server providers (CSPs) will account for 76% of cloud-related infrastructure hardware and software spending.*
What is hybrid HPC?
Hybrid HPC is the term that encompasses the solutions that address the three challenges many enterprises face:
Datacenter modernization: CIOs are rethinking their data center strategy by going off premises and handing over their data center operations to colocation and cloud providers. HPC growth often can’t be served by the space, energy, and environmental footprint of constrained legacy data center operations. This is why more companies are looking for ways to hand off their HPC operations, either through classic outsourcing or by hosting operations in a private or public cloud. A parallel HPC evolution is the move to the network edge. Often, the amount of data created at the edge is too large and time sensitive to be transferred to a remote data center. In many cases, the initial analysis and integration must be performed close to the data source at the edge of the network, such as in an autonomous vehicle, at the aircraft maintenance hangar or right next to the assembly line. This dual movement of HPC workloads—to off-premises and to the edge—requires integrated hybrid architecture from the edge to the data center to the cloud.
Agility: the digital platform use cases require rethinking what is expected from an HPC cluster. More than ever before, HPC clusters have to deliver simulation, analytics and AI insights to multiple Information System domains such as CRM manufacturing, logistics, field operations. Purpose-built infrastructure as a service (IaaS) and platform as a service (PaaS) solutions enable an agile, open and interconnected HPC infrastructure. For instance, HPC PaaS environments facilitate HPC application "portability" on premises, off premises and in the public cloud.
Financial risk mitigation: with the expansion of HPC across the product lifecycle, companies need models that allow for a reduction of upfront investments while helping to speed up deployment. Consequently, consumption-based payment models are becoming essential for on-premises deployments in the data center and at the edge. In this scenario, companies can leverage an on-site buffer to scale up or down on demand, paying only for the capacity they use.
Delivering a hybrid experience
HPE is committed to delivering a hybrid experience with these building blocks:
HPE GreenLake Flexible Capacity: companies need new financing models to reduce upfront investments and the financial risk, as well as the need for deployment speed induced by their ever-growing need for compute performance. HPE GeenLake Flexible Capacity offers consumption experience designed to help the customer manage and optimize their on- and off-premises clouds. In this model, companies leverage an onsite buffer to scale up or down on demand, paying only for the capacity they use. This flexibility enables them to achieve business outcomes including cash flow improvement, accelerated deployment and cost-effective capacity management.
The Partner Ready for Service Provider Program (PRSP): HPE has a network of CSPs such ScaleMatrix, Advania Data Center (ADC) and Opin Kerfi, that can provide colocation capabilities built to purpose HPC.
HPC IaaS and PaaS: HPE has developed a set of solutions aimed at providing a composable infrastructure built to purpose HPC. It includes a built to purpose HPC IaaS and the enablement of Linux Containers (today Docker EE).
HPE Data Management Framework (DMF) is the ideal solution to orchestrate the data movement from on-premises to off-premises to the public cloud.
The value of collaborative selling
The PRSP includes CSPs and some system integrators such as Accenture. With collaborative selling HPE improves the experience with additional capabilities such as HPC managed services (with TotalCAE), Singularity containers (Sylabs.io), Cloud Management Platforms (such Rescale and Atrio.io), and container engineering services with The UberCloud.
Where to start with HPE?
HPE has launched the Early Access Program (EAP) in 2018. Through the EAP, corporations gain access to demos, trials and support for proof of concepts to begin their Hhbrid HPC journey.
EAP projects start within HPE Centers of Excellence (CoE) dedicated to hybrid HPC. Projects often involve multiple HPE technology partners such as UberCloud.
HPE has co-invested with Advania Data Centers (ADC) to create its first CoE in 2016. This year HPE is creating another CoE with ScaleMatrix. Here's what they have to say about the experience.
“To me it’s clear that the future lies in the HPCaaS model, where simulation and optimization engineers can focus entirely on their product innovation while service providers like ADC take care of the whole HPC environment. I also think that we will see more and more companies looking to adopt a hybrid approach to HPC and AI, where they can combine their existing on-premise investments with on-demand access to resources in the cloud. As one of the only companies that can truly provide this hybrid service—and as the first HPE Center of Excellence dedicated to hybrid HPC. I feel that Advania Data Centers has a real competitive advantage.” - Guy D’Hauwers, Regional Director, Europe, ADC
“The HPC Center-of-Excellence (CoE) partnership with HPE is an especially exciting one for us here at ScaleMatrix. With our next-generation, DDC™-enabled colocation facilities as the backdrop, ScaleMatrix and HPE are able to provide clients with real-world examples of what the future of high-performance computing looks like today. The CoE for HPC will provide prospective users with access to HPE’s industry-leading Apollo HPC platform, deployed within ScaleMatrix’s revolutionary DDC high-density data center cabinets. When combined, DDC and Apollo offer one of the highest density compute deployments in the world, capable of supporting nearly any HPC workload with extreme energy efficiency and market-leading total cost of ownership (TCO). The HPE Center of Excellence will not only provide users with access to these powerful platforms for testing but will enable test users the additional benefit of being able to transition workloads to a hosted production deployment with little to no downtime.” -Chris Orlando, CEO & Co-Founder, ScaleMatrix
Unlike “classic” public cloud providers, HPE customers across many industries can customize their specific applications and workflows to drive outstanding performance, scale, consumption-based operations and economics from cloud technologies—and still benefit from built-to-purpose HPC infrastructure.
Ready to unlock hybrid HPC?
Email us firstname.lastname@example.org
*IDC estimates that worldwide whole cloud revenues will reach $554 billion in 2021, more than double those of 2016 and CSPs will account for 76% of cloud-related infrastructure hardware and software spending. Worldwide “Whole Cloud” Revenues Will Reach $554 Billion in 2021, According to a New Forecast from IDC—http://www.idc.com/getdoc.jsp?containerId=prUS43285217