Advania Data Centers has greatly invested in their Data Center and High-Performance Computing capabilities over the past years. These investments have resulted in a compelling value proposition for HPC operators, where the company can offer modular offerings. These offerings cover anything from HPC cluster hosting to application support and licence management.
Advania’s latest service is dedicated to HPC, where companies can outsource their HPC cluster to Advania’s HPCaaS cloud, hosted in Advania Data Centers in Iceland. The HPCaaS cloud can accommodate hosted HPC clusters, as well as allowing users to scale up and down as needed, while only paying for what they use.
The scalability has been offered by others in the industry for some time now, but on-demand resources have historically been expensive. Similarly, so has the cost of running HPC in a public cloud, whereas some HPC workloads are economical to run. Most, however, become way too expensive in the public cloud billing model, as well as some HPC jobs are not able to run in a public cloud because of technical or application licence issues.
Advania aims to address all these pain points with the new HPCaaS cloud, which has been developed in partnership with Hewlett Packard Enterprise.
“This expansion of our HPC platform is our next step to offer the HPCaaS cloud at scale to customers that demand increased flexibility.”
“We have developed a service and capabilities to address the needs of even the most demanding HPC operators, and it has proven to deliver great value to our customers,” Says Eyjolfur Magnus Kristinsson, Managing Director at Advania Data Centers. “This expansion of our HPC platform is our next step to offer the HPCaaS cloud at scale to customers that demand increased flexibility and want to be able to scale their environments as needed. While we are not disclosing how large this expansion is, I can say that this step of the expansion includes thousands of cores.”
This expansion includes the latest hardware from HPE, Intel and Panasas; purpose-built for HPC workloads with high speed interconnect and flexible storage options suitable for most HPC operators.
“I ran Linpack on 3200 cores and was able to achieve over 100TFlops, with a fantastic efficiency of 96%. I can state that our HPCaaS cloud performance is outstanding.”
“We are extremely happy to be able to expand our HPCaaS platform this quickly and have already signed the first customers on to the HPCaaS cloud,” says Hans Rickardt, HPC Specialist at Advania. “The Cluster is built with HPE Apollo 6000 chassis, and the servers are xl230a, configured with Broadwell E5-2683v4 2×16 core, 256 GByte memory, 1.92TB SSD, 10Gbit Ethernet and high-speed interconnect and 100Gbit Intel Omnipath. During cluster verification and burn-in test, I ran Linpack on 3200 cores and was able to achieve over 100TFlops, with a fantastic efficiency of 96%. I can state that our HPCaaS cloud performance is outstanding.”