The Ministry of Research, Science and Technology is looking for ideas on how best to spend $840,000, plus GST, on middleware for grid computing.
Grid computing, which harnesses the power of a number of remote computers to work collaboratively on large problems, is seen as an important facet of using the power of KAREN, the Kiwi Advanced Research and Education Network.
Grids tend to be more loosely coupled, heterogeneous, and geographically dispersed then clusters, says MoRST.
“The main advantage of grid computing is that each node can be purchased as commodity hardware which, when combined, can produce similar computing resources to a multiprocessor supercomputer, but at lower cost.
“Grid computing is enabled by the high-speed connections available through KAREN.”
However, effective and efficient operation of a grid requires a considerable layer of middleware between the operating systems of the individual computers and the demands of users. The $840,000 has already been earmarked for a grid middleware fund.
The fund could be used to develop a grid middleware strategy and/or an education programme “to accelerate awareness and acceptance of grid middleware, grid services, grid tools, workflow tools [and] federated databases,” as well as to licence appropriate software says MoRST’s request for proposal.
Part of it could also be used to establish grid partnerships with other nations and provide services and technologies of generic use to New Zealand researchers, such as “advanced wikis, collection management and social networking and science blogging tools”.
The technique of asking for proposals from industry is not new, says MoRSTs director of emerging technologies, George Slim.
“We do this often in areas where we don’t feel we have sufficient specialist expertise. Grid computing is an emerging and highly technical area.”
The ministry could have conducted a broad-ranging consultation among potential users to establish how best to spend the money, but such exercises tend to be long-winded and might not give optimum results, Slim says.
“Sometimes it’s better to get the people in the sector to put up their hands and tell us what they can do.”
But this is more than just a straightforward purchase of services; successful bidders will be asked to manage the fund, source complementary funding from elsewhere if possible and return any unspent money to MoRST at the end of the three-year contract.
The contract is scheduled to begin in February next year.