Sylabs was launched in 2015 to create a container platform particularly designed for scientific and excessive efficiency computing use circumstances, two areas that founder and CEO Gregory Kurtzer, says have been left behind within the containerization motion during the last a number of years. (For an evidence of containers, see this text.)
Docker emerged because the container of engine of selection for builders, however Kurtzer says the container options developed early on targeted on microservices. He says there’s nothing inherently unsuitable with that, however it ignored some forms of computing that relied on processing jobs as a substitute of companies, particularly excessive efficiency computing.
Kurtzer, who didn’t precisely simply fall off the open supply turnip truck, had greater than 20 years of expertise as a excessive efficiency computing architect working on the US Division of Power Lab, the place he based CentOS, an open supply enterprise Linux challenge and Warewulf, which he says has change into probably the most utilized stateless HPC cluster provisioner.
He determined to shift his consideration to containers when based Sylabs and launched the primary open supply model of Singularity in April, 2016. Even then, he had a imaginative and prescient of making a industrial model of the product. He noticed Singularity as a Docker for HPC environments, and would run his firm similarly to Docker, main with the open supply challenge, then constructing a industrial enterprise on high of it — simply as Docker had performed.
Kurtzer now desires to carry Singularity to the enterprise with a spotlight not simply on the HPC industrial market, however different excessive efficiency computing workloads resembling synthetic intelligence, machine studying, deep studying and superior analytics.
“These functions carry data-intensive workloads that demand HPC-like sources, and as extra corporations leverage information to assist their companies, the necessity to correctly containerize and assist these workflows has grown considerably,” Kurtzer wrote in a weblog put up saying the enterprise product.
Despite the fact that Singularity is designed to deal with completely different sorts of workloads, it nonetheless works with container orchestration instruments, particularly Kubernetes and Mesos, and additionally it is appropriate with Microsoft’s Azure Batch device and different cloud instruments.
Kurtzer indicated Sylabs presently has 12 staff, and is working on an undisclosed quantity of seed cash. It was funded by RStor, a startup itself presently working in stealth mode.
Featured Picture: Getty Pictures