Chapter 22

Computi ng is the science of information processing, and the generation and manipulation of information pervades all aspects of computing. Nevertheless, we feel confident in asserting that information—data and knowledge—is even more important in the Grid environment than in other domains. We make this assertion for two reasons. First, the virtual organizations that define Grid systems are formed to solve problems too complex or expensive to solve with local resources. Problem solving is concerned with the consumption and production of information. Thus, as we saw in the case studies of Part II, the generation, federation, and manipulation of information is the primary concern for many of the applications that motivate the development and deployment of Grid technologies, and is an important concern for all. The second connection is more subtle but equally important. The inherent complexity and dynamicity of Grid environments means that we cannot rely on implicit knowledge about system configuration, state, policies, and evolution. Instead, Grid middleware must be able to use and process information about the availability of services; their purpose; the ways in which they can be combined and configured or substituted; and how they are discovered, are invoked, and evolve. The Grid itself is a complex and information-rich environment. Thus, we are concerned not only with information on the Grid but also with information about the Grid. The two chapters that constitute Part V address this topic of information in a Grid context from the perspectives of data and knowledge management, respectively. Data and knowledge are vague terms: one person's data is another's knowledge , and thus we find some overlap in concerns and topics between the two chapters. However, they each address distinct and complementary parts of the overall problem.