Advanced Cluster Systems (ACS) hopes to bring grid computing into the business mainstream with its new
Math Grid Toolkit, a system for adapting Wolfram Research’s gridMathematica Workgroup Clusters.
Grid computing — using a network of computers to crunch numbers — has mainly been the purview of researchers, academicians and bioscientists. It can require some pretty extensive overhead in the form of cluster hardware and programmers to code applications that can distribute computational information out to multiple nodes.
Math Grid Toolkit lets users host modules developed for use with Wolfram’s Mathematica software using a workgroup cluster. Mac OS X-compatible, gridMathematica and Math Grid Toolkit run superbly on Apple’s Workgroup Cluster systems — portable racks of Xserve G5 hardware, explained Advanced Cluster System CTO Zvi Tannenbaum.
Math Grid Toolkit provides a hosting mechanism and a Web-based interface for running Mathematica modules on a gridMathematica Workgroup Cluster.
That liability in hardware, software and manpower is something that many businesses want to avoid, according to Bryan Minor, Ph. D., President of ScienceOps, a strategic partner of Advanced Cluster Systems. Minor told
that despite that issue, businesses can distinctly benefit from grid computing systems.
“Lots of businesses these days need to collect and analyze huge volumes of data in small amounts of time, not just biotech and aerospace” said Minor. “Data mining for companies doing Internet advertising, for example, statistical modeling, getting useful information from data sets that can be tens or hundreds of gigabytes in size — all of these tasks can benefit.”
ScienceOps develops algorithms used by ACS in its cluster solutions.
“We’re trying to tackle very large computational problems while reducing the overhead liability to companies,” said Tannenbaum. “Investors love this kind of business model.”
For customers who don’t already have a workgroup cluster installed, and who want to test the waters first, ACS owns its own high-power computing clusters that they can pay. Minor explained that this lets ACS’s client get their feet wet before committing to a large capital project.
“Some projects just won’t benefit from grid computing,” said Minor. “You can easily point to a lot of grid computing projects that have gone poorly over the years.”
“This system allows companies to lower their risk and lower their cost. It doesn’t cost nearly as much as setting a rack and hiring staff to write an application. It’s customized, and it’s scalable,” said Minor.
Customers interested in an MGT-based system are encouraged to contact ACS directly.