Biotech in Australia: equipment and services
The hardware side of the bio-IT market is mainly the domain of the big players--after all, you don't buy supercomputers from your local cut-price computer store.
According to SGI's Armitage, the main difference between first class science and other science is how fast the scientist can get through the "have data--ask question--get answer--ask new question" cycle. "The faster you can get the answers the faster you get the results," he said.
When you start talking about supercomputers, you start having to consider things such as vector versus scalar chipsets, single computers versus clusters versus grid networks, and comparing the conflicting claims of the hardware manufacturers, all touting the "fastest" supercomputer around.
Armitage claims SGI Altix 3000 computers have the advantage of a shared memory that is globally addressed. They offer a single system which consists of a large number of processors, giving the system a very fast interconnect. Meanwhile, IBM touts its Power4 chipset as one of the fastest chipsets for life sciences applications available today.
The world of the supercomputer is highly confusing to anyone not intimately associated with it, according to Philip Tannenbaum, manager of the Joint Bureau of Meteorology/CSIRO High Performance Computing and Communications Centre (HPCCC), which recently upgraded its weather forecasting supercomputer. He told ZDNet Australia that supercomputers normally only attain a low percentage of their advertised peak performance.
"The other model that life sciences organisations use is Linux clusters," said IBM's Palanca.
Clusters have the benefit of creating a lot of processing power for relatively little cost, but SGI's Armitage is skeptical that this is the best method. "A lot of people who have purchased clusters for major science applications tell us they are spending 25 percent of their time looking after the clusters. They're not doing science, they're doing computer science," he said. "There'll be no Nobel prize for an institution if scientists are spending 25 percent of their time on computer [maintenance]."
Another system used for high throughput computing is grid computing, which parcels out portions of a heavily number-crunching project to a large number of different computers over the Internet.
Grid computing has definite advantages, being cheap and (theoretically) having limitless processing power. However, apart from finding people to donate "computer time" to the project, grid computing can run into trouble authenticating data, as SETI@HOME discovered last year.
Of course, not everything in a biotech company has to be high-end. The Datamonitor survey revealed that areas most likely to be outsourced by large life science companies are desktop maintenance, hardware maintenance and procurement. Levels of outsourcing in the bio-IT arena are set to increase, according to Datamonitor, but biotech companies are skeptical of claims made by vendors as to what their products can do and how much money they can save.
On the software front, since most tools commonly used were originally developed at universities, it's usually open and freely available.
The "free" mentality makes selling software quite difficult, said Emphron Informatics' Thomas.
"A lot of the work is in the public sector and that means it has a certain character," said Thomas. "Fairly restrictive budgets, an emphasis on research grade software rather than production grade software...What we don't see very much in Australia is enterprise-scale bioinformatics systems. What we see a lot of is bioinformatics applications targeted at relatively small research groups, and what we don't have in Australia is substantial enterprises with vast data resources requiring expensive infrastructure to keep them working."
Meanwhile, the services segment could well be a major growth area in IT spending by life science companies.
Market research firm Gartner has predicted that by 2005, one third of IT revenue from life science companies will derive from services and solutions for the research and development process--which is a marked shift from a few years ago when the majority of investment was in enterprise applications such as finance, human resources, supply chain management, CRM, drug marketing and pricing.
And with Australia's biotechnology industry tipped to exceed US$5 billion, perhaps this "new" area will prove to be the country's next leading growth engine.
| Intro |
 |
|
State of play |
|
|
Life and science |
| |
Jobs galore |
| |
Equipment and services |
