Excel provides powerful tools to analyze, communicate, and share results. It is widely used for pricing derivative and other financial products.

XcelaGrid(TM) makes it possible to seamlessly exploit the capabilities of grid computing resources from within the Excel environment. The term grid computing refers to the massive integration of computing resources to achieve a level of performance unattainable by any single machine. Computational grids are persistent networked environments that integrate geographically distributed computers, databases, and services. XcelaGrid offers the best of both worlds: the rapid analysis/design/prototyping capabilities of Excel and the high performance backend offered by a grid.

XcelaGrid™ provides a toolkit for analysts, and other users of Excel to use the power of grid computing without having to master the complexities of grid computing. They will be able to locate resources and run their grid related workflow from within their Excel environment.

Problems that lend themselves to grid computing using XcelaGrid include:
• Distributed analysis of large amounts of data: A large amount of data is often generated in geographically disperse locations, and then sent to a central location for analysis and visualization. Using grid technology it is possible to analyze the data at the source, and possibly even generate visual representations of the data. The analyzed data and the visualizations may then be sent to a central location for assembling the overall view of the data. This not only reduces the computational load at a central facility, but may also lead to a reduction of data being transported over the network.
• Task Farming: Certain problems are not amenable to analytical solutions and cannot be solved by simple computer algorithms. These need massive amount of CPU power not available in the single machine. Brute force methods such as Monte Carlo methods are often used to tackle such problems. These methods are often readily parallelizable to exploit the power of grid computing.
• Distributed multi-component simulation: In this situation, it is desirable to construct a large, simulation where the constituent elements each reside on different computing systems. The computational burden of all simulation models considered together may be such that it is not feasible to run all of them on a single machine. A preferable approach is to coordinate the distributed operation of each simulation component from one location.