Computational Grids are persistent environments established by combining diverse high-performance computing systems and specialized peripheral hardware. The ultimate goal of grid initiatives is to provide ubiquitous access to networked resources comparable to the omnipresence of the electrical power grid. Access to the grid must be offered via a transparent, unified view across individual systems, organizations and even international borders. In order to exploit the processing power provided by grids, software applications must be adapted and tailored to its characteristics and concepts. Since grid computing is in many aspects a generalization of technologies that were originally developed in the context of high-performance computing and networking, it is an essential fact to employ parallel and distributed computer systems and techniques. Consequently, grid computing affects every aspect of program development, which in turn must be addressed by parallel and distributed programming tools and environments. This includes tools for monitoring, testing and debugging, performance analysis and tuning, load-balancing and scheduling, and software visualization.