ROCKVILLE, Md.--Celera Genomics, the new unit of Perkin-Elmer that aims to complete a sequence of the entire human genome within three years, has selected Compaq Computer to design and maintain its IT infrastructure, based on Compaq's Alpha architecture. Celera President Craig Venter has said he intends to make his company the "definitive source of genomic and medical information."
The core of Celera's IT operation will be a mixed environment using Alpha-based servers and Alpha and Intel workstations running 64-bit Digital UNIX and Windows NT, according to a Compaq spokesman. Officials from Celera, which is in a mandated quiet period, declined to comment on the partnership. Compaq said it will initially deploy Digital AlphaServer 8400's and Digital AlphaServer 4100's, as well as a number of Digital AlphaServer 800's and 1200's running both operating systems. The servers will be complemented by approximately 200 Compaq Professional Workstation 8000's for research use. High-performance storage systems based on Compaq's StorageWorks technology will also be implemented, Compaq said.
The combination of a very fast microprocessor and the memory capabilities of an Alpha system supporting Digital Unix was appealing to Celera, according to Bill Blake, Compaq's director of high-performance technical computing. The difference between 32 and 64 bits can exceed a 10-to-one clock time, he noted. "The work that gets done in one hour on Alpha can take in excess of 10 hours on other systems," Blake claimed.
He added that Compaq's experience providing internet search tools caught the attention of Celera executives. "Sixty-four bits helps with the speed of acquiring and assembling information," Blake told BioInform, "but beyond that is the notion of how to make that amount of information accessible to a lot of people."
Compaq will provide Celera with its Alta Vista indexing and search capabilities that will allow the company to use the internet and electronic commerce technology to distribute its findings to the scientific community. Explained Blake, "With Alta Vista we put on the internet a capability that well over 30 million people access daily. So you can think of a single application of that which accesses, instead of a database of all the websites in the world, the biggest biological database in the world."
To support the Celera infrastructure, a team of Compaq engineers will be stationed at Celera's facility here. "We have put together a team that will cover the whole gamut from keeping the machines running and building the network infrastructure to helping with the application development," Blake said.
Although most Compaq engineers will be based at Celera short-term "to get the infrastructure built, hook up the network, and get the systems running," Compaq's relationship with Celera's product development group will be "multiyear," according to Blake. "Probably half a dozen people will be working with them on things above and beyond just getting a large system in place," he added.
Blake noted that Compaq stands to gain more than just a new customer from the deal. "For us it's a perfect opportunity to really learn how to develop a system that scales up to this level of performance. It's important to us to find the leader in a given market or industry segment and work with them very closely. We get extremely valuable information on what will be needed three or four years down the road."
With the human genome's 6 billion characters, together with its associated annotations, Celera will require a multiterabyte database. The operation will ultimately consist of a cluster of machines with about 75 terabytes of storage, according to Blake. "This is performance that we typically only see in national labs and defense related projects," he said. "From a private sector, life-science point of view this is an incredibly interesting and powerful use of high-end computing."
Coupled with what he described as "an intriguing connection to the internet," Blake said the Celera operation is "a wonderful showcase of the kind of large-scale science projects that will be going on in the next decade."