Grid Computing for Electromagnetics

In the previous section, we described the nature of the four main building blocks of an environment suitable for CAE of aperture arrays. From now on, we refer to the four building blocks by introducing the following acronyms:
Evaluation of the horns' scattering matrices: analysis of the feeding system (AFS);
Evaluation of the aperture array's scattering matrix: analysis of mutual coupling (AMC);
Evaluation of the scattering matrix at external physical or electrical ports: evaluation of the scattering matrix (ESM);
Evaluation of the radiation pattern (ERP).
As discussed earlier, some of the numerical tasks are extremely heavy from a computational point of view. Both the mutual coupling of apertures and the evaluation of the generalized scattering matrix of the whole circuit are complex. The former is due to integration problems; the latter is due to the implied matrix algebra, with several large matrix inversions see (5.20). Moreover, the four building blocks are based on rather different formulations and approaches and can often be developed in the framework of large projects by different research groups with different software methodologies. Consequently, the effective implementation of an environment for CAE of aperture horn arrays must basically face two problems: the need to sustain high computing performance and the need to allow cooperative engineering. In this section we demonstrate how parallel computing can solve the first problem. In the remaining part of the chapter, it is demonstrated that GC, while supporting cost-effective parallel computing, allows a complete and secure solution...