Secant Technologies, Cleveland, Ohio, previewed the Linux-based compute farm application it is building in cooperation with Incyte Genomics and IBM at the recent LinuxWorld Conference and Expo held in New York.
The system, which the trio of companies began building in September, is intended to be a core element of Incyte’s Genomic Knowledge Platform for data integration, which will also incorporate IBM’s DiscoveryLink data management software.
The collaboration marked the first partnership of IBM’s well-publicized $100 million life sciences initiative, but it was also Secant’s first partnership in the genomics market. Paul Horan, vice president of business development for Secant, said that the company is “looking to aggressively focus on the life sciences market,” and plans to announce several new partnerships with life science companies over the next few months.
“Both from a technical standpoint as well as a sales and marketing strategy, we’re focusing in the life sciences area,” said Secant CTO John Pompeii.
Pompeii said that Secant’s cluster management software was initially developed to handle smaller clusters than Incyte’s — hundreds of processors as opposed to thousands — so they had to modify the software to handle the larger scale. He said the system now has the potential to support clusters of up to 10,000 machines using a design that builds the clusters as groups of subnetworks.
New capabilities that Secant is bringing to Incyte’s current Linux farm include increased scalability and the ability to partition the farm into computing regions that allow particular areas to be reserved for higher-priority jobs.
“It adds a management layer that allows for an organization to very easily partition the farm up and use it in various ways instead of having to deal with it all as one large unit,”
Pompeii said. Pompeii said that Secant is also adding a suite of management tools that will allow the farm to be monitored for performance. A centralized database in the new farm contains plans for the complete configuration, information that was previously required on each individual node.
“That allows for actual computers to be added and removed from the farm while it’s up and running live,” Pompeii said.
Horan noted that the LinuxWorld demonstration, which showed how to create, submit, and execute computational jobs and manage the compute farm’s resources, was only “a technology preview to show what was possible.” He added that the final details of the terms of the three-way collaboration “are still being hammered out.”
Incyte previously said that initial products for the Genomic Knowledge Platform would be released in first-quarter 2001. An Incyte spokesperson could not be reached for comment on the current status of the project.