Linux server and software maker Penguin Computing Tuesday signed paperwork saying it will acquire privately held Scyld Computing, best known for developing Beowulf high-performance clustering software.
Under the terms of the agreement, San Francisco-based Penguin will issue shares of its stock in exchange for all of the outstanding securities of Annapolis, MD-based Scyld. Sam Ockman, founder and CEO of Penguin Computing, said he will lead the combined companies as president and CEO. Donald Becker, former NASA scientist, early Linux networking developer, founder of Scyld, and key developer of the Beowulf software technology, will serve as the Chief Technology Officer.
“Over the past four years, Scyld has invested aggressively in the development of its software and has earned a strong reputation in the Beowulf clustering market,” Ockman said in a statement. “The Scyld software reduces the complexity of provisioning and managing a Beowulf cluster to that of administering a single Linux workstation.”
While Penguin said it would continue selling its Linux-based servers, it is offering customers the ability to “choose from a wide array of highly-configurable server building blocks, advanced cluster management software and a wide range of technical consulting services”
“For most applications we’ll soon be able to offer an extremely compelling alternative to traditional monolithic supercomputers,” Becker said.
Scyld said it will maintain its own brand as a software division within Penguin Computing and will continue to operate from its East Coast office. Executives with both groups said they don’t anticipate any layoffs at either company.
Beowulf is a cluster of Linux systems which emulates a supercomputer high-performance massively parallel computer built primarily out of commodity hardware components, running a free-software operating system like Linux or FreeBSD
Scyld said its Beowulf operating software allows for a single system image for cluster process management. Processes running on cluster “Computation Node” computers are managed from the cluster’s front-end “Master Node”.
Processes start on the front-end node and migrate to a cluster node. Sclyd said process parent-child relationships and UNIX job control are maintained with migrated tasks. Cluster compute nodes are not required to contain resident applications. Their hard disks are used for application data and cache. The company said this approach eliminates version skew common with previous generation clusters.
Penguin almost exclusively uses Intel for its large-scale boxes including Xeons, Pentium III and Pentium 4 chips for its high-end Relion-class clustering server and Nevieus workstations. The company defers to AMD
Athlons for its less expensive models.
Both companies have received support from the venture capital community. Lead investors in Penguin Computing include ING Capital and Dragon Venture. Bob Young, co-founder and former Chairman and CEO of Red Hat
, is an investor in Scyld.