"In the past, when developers referred to building a computing grid, they primarily had the hardware and operating systems level in mind. Grid computing was all about brute force -- tying together as many machines as possible -- to crunch through computationally intensive tasks like financial market simulations, or oil and gas exploration.
"Today, the focus of grid computing development has moved up a step, to the application level. While sophisticated hardware linkages aren't exactly commonplace, they're much better understood than they were in the past, and form the basis for many commercial platforms available today. The next challenge for enterprises is to integrate heterogeneous systems and applications into a wider-ranging logical grid, one that can bring new performance highs to critical enterprise applications.
"While enterprises can always build their own middleware, linking this wide a range of devices, managing device-to-device security, and scheduling and coordinating jobs between devices with different levels of computing power can pose thorny problems for grid developers. At minimum, developers at companies interested in an application grid face a steep learning curve..."
