Article by Naoki SHIBATA, CEO and High Performance Cloud Architect
Those who have never used HPC in the cloud often cite connecting to the cloud and monitoring usage fees as major sources of concern. Recognizing this, we have utilized technology to address these concerns and enable clients to feel secure building and utilizing their first cloud-based HPC cluster.
How Do I Connect to the Cloud?
If you’ve never used cloud services, you might be wondering how to reliably connect to the cloud from your corporate LAN, or what exactly you should do to create a connected environment. No matter how well thought-through your HPC cloud solution may be, it will not be worth anything unless you can connect safely and comfortably to your cloud environment.
XTREME-D’s flagship product, XTREME-DNA, is a web-based offering that provides high-end, cloud-based supercomputing, developed by highly skilled HPC architects. XTREME-DNA was designed with the user in mind, so its UI/UX is visual and easy to use, and it walks the user through providing the parameters required to configure an optimal computing environment for their specific needs. Users can quickly access their own ready-to-use system, with OS, middleware, job scheduler, and configurations, all set up and ready to go in as little as 10 minutes.
And if you require even more hand-holding, XTREME-D offers technical support and consulting to help users construct their ideal cloud environment, connect to the cloud, and immediately start running jobs. While not part of the “official” product specs, XTREME-D’s excellent customer support comes with every instance of XTREME-DNA, and goes a long way in creating customer trust. This deep level of technical support is possible only with many years of building cloud environments for customers.
How Do I Monitor Cloud Usage Fees?
HPC in the cloud consumes a lot of cloud resources. And charges start accruing from the moment the power is first turned on, due to usage-based billing. For this reason, a real-time record of consumption cost is key, and XTREME-DNA provides that for you via a sophisticated dashboard that lets you track that data at a glance.
As an example, if you utilize a 1024-core HPC system for 24 hours, it consumes about $5,000 in cloud usage fees. Given this base point, just imagine your surprise should you forget to turn off your HPC cluster, and only find that out by looking at your monthly invoice many days later!
With XTREME-DNA’s budget control features, you can use only what you need for a fixed monthly fee. XTREME-DNA provides tools to configure monthly budget, set up alert thresholds, and obtain detailed usage reports, so that you can avoid overuse entirely. And after deployment, it is extremely easy to manage supervision of your clusters, by turning on or off or deleting your virtual supercomputer with the click of a button.
Delivering the Ultimate HPC Cloud Service
With the built-in ease of configuration and budget control features of XTREME-DNA, you can use cloud HPC with confidence and not worry about accidental cloud usage and the resulting unexpected budget expense.
But XTREME-D wants to go even further in delivering the ultimate HPC cloud service. In the public cloud arena that most users utilize (AWS, Azure, Google Cloud Platform), it is complicated to get virtual machines to maintain HPC performance. XTREME-D is currently working on next-generation solutions that remove this complication by an innovative utilization of infrastructure.
The full concept will be announced shortly, so please stay tuned for the next newsletter, and watch this space!