Joe W., Really like this one. A couple days ago there was that INRG piece about that UK genome director. We haven't talked about storage needs in this arena.....
April 10, 2001
The 800-Pound Genome
Decoding the human genome wasn't the culmination of biocomputing - it was just the end of the beginning. The sheer horsepower and storage demanded by the analytic and database challenges of bioscience applications can swallow information systems whole and beg for more.
IT vendors that are reducing expectations in other areas might want to consider the burgeoning life sciences market, an initial forecast for which is presented in Figure 1, drawn from IDC Executive Insight The Convergence of Biology and IT: Market Impacts of the Bioscience Era (IDC #24168, March 2001).
Figure 1 - Biosciences IT Market Revenue, 2000-2004
Consider Celera Genomics, the private company whose methods claimed first prize in sequencing the human genome. Celera's infrastructure uses:
$1 million of electric power annually
70TB for databases, growing by 15-20GB per day
1.7TeraFLOPS of aggregate processing power
900 Compaq AlphaServer processors
Six Paracel GeneMatchers (each with thousands of processors)
Paracel is a massively parallel processor (MPP) manufacturer that Celera purchased last year.
Big Science Just Got Bigger
Biocomputational requirements now exceed those of other "big science" applications, such as weather forecasting, subatomic physics, and nuclear weapons research. In addition to genomics, disciplines such as bioinformatics, pharmacogenomics, proteomics, and computational chemistry are hungering for more powerful systems.
J. Craig Venter, Celera's president, has noted that "just three years ago, the computational needs of biology were thought to be minor and irrelevant to the computing industry. Today, biologists are setting the pace of development for the industry."
Venter's remarks were made during the announcement of a cooperative R&D agreement between Celera and Sandia National Laboratory that will develop a supercomputing platform for mining and modeling genetic data. The $25 million project will be powered by a dense fabric of Compaq AlphaServers and is expected to provide 100 trillion operations per second (TeraOPS) of raw compute power, making it the world's largest supercomputer.
Centers of Gravity
The biocomputing universe is still expanding and has yet to condense into recognizable vendor constellations.
Compaq Versastor anyone?has drawn a bead on the technical computing market and achieved high penetration of the expanding life sciences segment. Its focused investments in solutions development (e.g., staff hiring and training, third-party software partnerships, and requirements analysis) seem to be paying off.
IBM's large orbit around scientific, academic, and government technical computing isn't decaying, however. According to www.top500.org, IBM will soon have built six of the top 10 supercomputers and 215 of the top 500.
NuTec Sciences (a bioinformatics solutions provider) has tapped IBM to build a 7.5TeraFLOPS supercomputer, in partnership with U.S. National Human Genome Research Institute. The cluster will connect 1,250 IBM e-Server p640s running IBM's DB2 Universal Database, supported by a NuTec operating system, 2.5TB of memory, and 50TB of online storage. NuTec will use it to manage, mine, and integrate genetic data and share this information with the life sciences community.
Other technical computing vendors, such as Hewlett-Packard, NEC, SGI, and Sun, haven't scored biocomputing wins as effectively as Compaq and IBM but are expected to aggressively seek larger shares. Many outcomes are possible, as the field is far from mature and its ecosystems are still forming.
Life Lessons for Biocomputing Vendors
The emerging biocomputing market needs:
Hardware. High-end servers, MPPs and other supercomputers, networking gear, workstations, and instrumentation
Storage. On the order of petabytes, overall
Services. Hosting and application provision services for decoding genomes and analyzing their functions
Software. Knowledge management, database systems, and applications for informatics, sequencing, proteomics, computational chemistry, visualization, and more
Services and storage will become increasingly important over time, as Figure 2 displays.
With its IT infrastructures and business models still immature, limited supplies of relevant IT expertise, and overwhelming knowledge management requirements, biocomputing's solution space is underpopulated. Current software suppliers (mostly from an academic-based cottage industry) and service providers can't keep up with biocomputing's pace of change and scale of problems.
Even though biotech is not feeling as much economic pain as high tech recently has, its continued growth still faces obstacles. Investment sources, smarting from other losses, are especially finicky and may desire returns more quickly than can reasonably be expected. Also, a plethora of knotty policy and ethical issues that can't be wished away portend uneven progress for the industry as a whole.
Still, in the current economic climate, it is hard to find a more robust IT market than biocomputing.
Figure 2 - Biosciences IT Market Segmentation, 2000 and 2004
--Geoffrey Dutton
Related Research
Take a look at the graphs here...
idc.com
Sure enough while the whole pie gets bigger storage overhauls servers sometime over the next 3 years. Compaq won't have any trouble selling virtualization to theis market.
Open Q) Over on Yahoo there has been talk about BRCD and CPQ. With the pr piece last wk I didn't see any hint that
there was big trouble between the two. That would suggest CPQ would dump them. But the future doesn't wait for anyone, we are shifting to the time of directors and that is an opportunity for those that have products in that categry already. I think INRG earnings are also the 19th and while all others reporting that day have issued est. INRG hasn't said anything. INRG will be the bellweather for director sales. |
