Hybrid HPC can deliver flexibility and scalability at a low TCO. Learn how cloud provider ScaleMatrix developed a Center of Excellence to reap these benefits.
That's why the company, which provides cloud, colocation, and other services to the western, southern, and eastern US, recently decided to implement hybrid HPC by creating a new Center of HPC and Cloud Excellence with the help of HPE's Early Access Program (EAP).
The Center of Excellence program was developed to help IT departments learn and incorporate new and developing technologies through hands-on activity and to act as a sandbox for enterprise clients. Thanks to guidance from HPE technologists, ScaleMatrix is now up and running with hybrid HPC and already enjoying results.
The hybrid HPC advantage
HPE's hybrid HPC cloud solution was created to support organizations that want to high-performance data analytics into their operations. It enables enterprises to use both onsite data centers and private or public clouds for an optimal combination of speed, security, agility, and ROI.
The biggest advantage is that it provides the best of both worlds. On-premises systems offer the speed needed to rapidly provide data to the organization, while cloud-based deployments expand storage resources exponentially. In addition, hybrid HPC can function on both public and private clouds, providing organizations with more choices for highly customized cloud configurations.
Another advantage is a lower overall total cost of ownership (TCO) due to both those customized configurations and HPE's consumption-based payment model, which allows you to pay only for the cloud storage you use.
A seamless onboarding process
This past year, ScaleMatrix announced that it was launching the first North American Center of Excellence (COE). The new COE will be located in the company's San Diego Data Center flagship and will offer a specifically designed HPC infrastructure enabled by ScaleMatrix' Dynamic Density Control™ (DDC™) platform. According to ScaleMatrix, its DDC platform, which includes a liquid-air hybrid cooling platform, allows for up to 52kW of workload and is 30 percent more efficient than competing data center cooling methods.
HPE provided ScaleMatrix with a list of steps it could follow to ensure that the process was as smooth and efficient as possible. After all the agreements were signed and the documentation supplied for a company in the EAP, the scope of the work was developed, the amount of investment necessary was determined, and a compute and storage tenant was provisioned using HPC IaaS so the trial can commence.
At this point, a starter kit was then supplied by HPE, the hardware setup was implemented, and the software was installed with the help of HPE advisers. The starter kit used for the ScaleMatrix' COE included an Apollo configuration, HPE Performance Cluster Manager, Docker EE, and HPE HPC IaaS 1.0. The IaaS, which is available via a Pointnext contract, was built especially for HPC and has a multi-tenant composable infrastructure. Since infrastructure-as-a-service (IaaS) is one of the main categories of cloud computing services, deploying the HPE HPC IaaS enables cloud providers to get to market faster and reduce development costs.
Setting up the configuration
ScaleMatrix used a two-rack configuration with one shared-management control plan to deploy, image, and monitor both racks. The left rack cluster is equipped with Intel® Omni-Path Architecture (OPA), enabling the system to scale to petaflop levels of storage, while the right rack cluster provides Ethernet for OpenStack or HPCM and HPC IaaS. Authentication is provided by the open-source Keystone software.
The results should provide ScaleMatrix customers with improved TCO and system performance by allowing them to offsite their data center operations while keeping time-sensitive operations at the edge of the network. In addition, the built-in agility of the IaaS solution will adjust to whatever circumstances are required by the HPC cluster, and because HPC is a consumption-based payment model, it is extremely cost effective. The results?
“With our advanced DDC-enabled colocation facilities as the framework, ScaleMatrix and HPE can deliver test environments for HPC workloads to push clients’ performance and see the value of the Center of Excellence.”
-Chris Orlando, CEO, ScaleMatrix
Learn more about the hybrid approach to HPC
If you are running a data center and wish to optimize your use of hybrid HPC, consider joining HPE's Early Access Program and creating your own Center of Excellence.
To discover how HPE can advance your organization’s HPC infrastructure with Hybrid HPC resources and solutions visit our Cloud28+ page or contact an HPE sales representative by emailing hpc-eap@hpe.com
TechExperts
Our team of HPE and other technology experts shares insights about relevant topics related to artificial intelligence, data analytics, IoT, and telco.
