An Oracle-led life sciences industry group, a pharmaceutical customer win for IBM, an academic customer win for Sun, and several new product announcements highlighted last week’s activity among IT vendors jostling for position within the life sciences marketplace.
In one particularly unabashed display of the IT sector’s infatuation with the life sciences market, Cray opted to describe the data transfer rate of its recently launched Solid State Disk memory system in terms of “human genomes per second” last week. According to Cray, the new 224-gigabyte SSD system “can hold 27 copies of the human genome and transfer data at a rate equivalent to 100 human genomes per second.” The company kindly translates that into 80 gigabytes per second for its high-end computer customers who may not be working in genomics.
But Cray’s not the only company seeking to woo life sciences customers, a fact borne out by last week’s events:
Oracle Partners with Life Science Domain Veterans
At its April 9 “Oracle Life Sciences Day” — the second such event the company’s held since December — Oracle announced a new platform dubbed the Information Architecture for Life Sciences.
Built upon the E-Business Suite the company built for its dot-com customers to provide a single data schema to consolidate data, the architecture offers two new features. The first is the general “information architecture” that opens up the API of the E-Business Suite. Additionally, the platform’s been tweaked a bit to handle the demands of biology, according to Jon Simmons, vice president of life sciences for Oracle, who told BioInform that the company has spent “the last couple of months” working with customers and partners to “extend that data schema for science objects.”
Working with 10 partners including Sun Microsystems, PriceWaterhouseCoopers, MDL, GeneticXchange, Daylight, Dendrite, and Acero, Oracle has put together a platform for both scientific and business applications of interest to biotech and pharmaceutical companies, Simmons said. Partners have agreed to make their offerings part of an “application cluster environment” to consolidate scientific data before it moves downstream into clinical and even sales and marketing environments.
Melanie Flanigan, a spokeswoman for GeneticXchange, said the company is acting as Oracle’s primary partner for data integration in the life sciences market. “We can deal with very complex data sources out there in the life sciences industry that are different than relational databases,” she said. GeneticXchange will provide “quick loading” of data types and structures that are outside of Oracle’s capabilities into the Oracle environment.
While partners stand to benefit from exposure to Oracle’s strong customer base, there are no revenue-sharing aspects of the partnership strategy.
Merck Picks IBM for R&D
IBM has computer systems running the business side of many pharmaceutical companies, but it hasn’t had much of a presence in research informatics. This makes Merck’s recent decision to put an IBM supercomputer cluster to use for drug discovery research “a significant win for IBM,” according to a company spokeswoman.
Merck will use five 32-way p690 systems with a total of three terabytes of disk storage for basic research tasks, including database searching, chemical property computation, quantum mechanical computations, and clustering. These servers, which will be located in New Jersey, Pennsylvania, Canada, and the UK, join two IBM SP supercomputing systems that are already installed at Merck.
Sun Servers to Power KEGG
Sun Microsystems said last week that the Bioinformatics Center of the Institute for Chemical Research at Kyoto University has purchased three Sun Fire 15K servers — each configured with 72 CPUs, 144 gigabytes of memory, and 15 terabytes of storage — to support the KEGG (Kyoto Encyclopedia of Genes and Genomes) database as well as other projects.
Minoru Kanehisa, director of the Bioinformatics Center, told BioInform that one of the Sun Fire servers would be a development machine for the KEGG project, while the other two are service machines to support the GenomeNet online service (www.genome.ad.jp), which includes KEGG.
In October, the ICR purchased two SGI Origin 3800 systems, which Kanehisa said are reserved for “computationally intensive tasks” such as computations for the SSDB (Sequence Similarity Database), as well as Blast, Fasta, Motif, and other searches.
Prior to the new Sun servers, KEGG ran on a Sun Enterprise 10000, which Kanehisa said was used both for development and service and provided only 16 CPUs, 2 gigabytes of memory, and 500 gigabytes of disk space.
Kanehisa added that Sun’s SMP (symmetrical multi-processing) architecture was a selling point. “All the KEGG objects are represented and computed as graph objects,” he said. “In order to process a huge graph object, such as a protein interaction network in the living cell or the universe of all genes in all known organisms, we need a large SMP architecture.”
Sun, IBM Battle Over Server Market
Sun and IBM pulled no punches last week, openly attacking each other in their respective midrange server launches.
IBM introduced its p670 family of servers built on its Power4 microprocessor technology. Noting in a statement that, “the new server intensifies the contest between the latest-generation Unix technologies from IBM and Sun,” IBM said its systems would provide up to a 34 percent cost savings over comparable Sun machines.
According to IBM, its eServer p670 line would range from $178,270 for a 4-way system to $535,522 for a 16-way system; compared to $250,457 for a 4-way Sun Fire 4800 and $776,795 for a 16-way Sun Fire 6800.
Sun blasted back the following day with the launch of its Sun Fire 12K server, which it has dubbed the “Regatta killer” in a bid to compete with IBM’s p690 Regatta system. The 12K is positioned between the company’s Sun Fire 6800 and Sun Fire 15K and uses the same Uniboard CPU/memory board found in all Sun Fire systems.
With a $500,000 price tag, the new server “offers technology that is years ahead at fractions of the price of IBM’s p690,” according to the company.