Grid computing's evolution can best be measured by how it compares to electrical grids, concludes a new paper by Madhu Chetty of Monash University and Rajkumar Buyya of the University of Melbourne.
Much has been made of the utility computing vision for Grid computing, where compute power would be delivered just like electricity from an outlet, without the user giving a thought to where the power came from or its complexity and reliability. In an article published in the July/August issue of the IEEE journal "Computing in Science and Engineering," Chetty and Buyya took the metaphor a step further with a point-by-point comparison of computational and electrical grids.
One obviously similarity between computational and electrical grids is that both aggregate heterogeneous power sources and deliver them seamlessly, be it thermal, hydro or nuclear power, or PCs, servers or supercomputers, they said.
In the 200 years since Alessandro Volta invented the battery, the electrical power grid has become "one of the most advanced and evolved grids in existence," Chetty and Buyya said. Computational grids, on the other hand, are "a new and emerging field, now in a state in which the electrical power grid was almost a century ago," they wrote.
"The use of computational grids for solving real-world problems is still limited to research labs and a highly specialized scientific community funded by government agencies," they said. "Pushing grids into mainstream computing will require major advances in grid programming, application development tools, application- and data-level security, and grid economy."
The comparison to electrical grids found computational grids lacking in several areas. "The need for an operational model (a regulated system or otherwise), proper division of the computational grid into regional pools, coordinated system operation to ensure network stability, and ease of use must all be priorities in further grid development," Chetty and Buyya wrote.
Buyya investigated Grid economy concepts in his
Ph.D. thesis, and the University of Melbourne's Gridbus Project is building on that work to develop
end-to-end quality of service-based utility computing.
The article can be found at: http://www.buyya.com/papers/WeavingGrid.pdf.