Cluster Resources has spent the better part of the decade working on high performance computing (HPC) automation. But now Cluster has announced it’s expanding its Moab unified intelligent automation middleware technology into the datacenter.
As part of the move, the company is also changing its name to Adaptive Computing and creating a second division to service the datacenter market.
Adaptive’s Moab software is widely used in high performance computing, including 12 of the 20 fastest computers on the Top 500 supercomputer list.
Normally, HPC systems are considered a different animal than the standard issue datacenter, but Adaptive Computing sees the two coming together.
“The foundation for datacenters for the enterprise is starting to look a lot like the architecture of supercomputing facilities. That’s creating an opportunity for ourselves,” Peter ffoulkes, vice president of marketing for Adaptive Computing told InternetNews.com. “The capabilities you need to automate a datacenter is exactly what we’ve been doing alongside IBM and HP for years.
Effective today, Adaptive has two business units: Cluster Resources, which will continue to serve HPC customers, and Adaptive Computing, which will serve the datacenter market. For the datacenter market, Adaptive is introducing the Moab Adaptive Computing Suite platform, which adds functionality specific to the datacenter.
The Moab series of products is built around a middleware technology that manages workloads to match demand. Moab Adaptive Computing Suite (MACS) works in an IT datacenter, where the usage is different. A supercomputer like Roadrunner at Los Alamos National Laboratories is used to run massive simulations that can utilize every core.
A datacenter, on the other hand, might have applications assigned to each physical processor. One server might run ERP, another runs HR, a third handles mail. So with the workloads distributed and scattered, MACS had to be tuned differently.
Managing virtual machines alongside physical resources
MACS adds support for communication between enterprise apps, support for transactional pipelines and lightweight apps that are often implemented in a virtual machine environment. MACS also have the ability to manage virtual machines alongside physical resources.
“That means when making priority and policy decisions about what workloads to place where and when, you are now able to think of resources as just this virtualized pool of logical resources. So whether an app is deployed on a single physical resource, a group of physical resources or as part of a virtual machine environment inside of a server, it’s just a part of the calculations we take into account,” said ffoulkes.
In addition, MACS has adaptive capabilities where it monitors requests coming in to the network and can reconfigure the environment. As the need for resources change from day to day, if there’s a different mix of requirements, such as needing more compute power for certain tasks, MACS shifts resources around to give the priority applications all the horsepower they need.
The Moab Adaptive Computing Suite is available now from Adaptive Computing.