This survey of research on data management for wireless applications focuses on new approaches to providing database services in an environment where the client (for example, a smart phone or personal digital assistant (PDA)) is moving around in a wireless cellular infrastructure.
The book starts with a history of, and tutorial on, wireless networks. The author describes the architecture of wireless networks (base stations, home location registers, and so on) and some of the problems that wireless networks have to solve. Both personal communications service (PCS) and global system for mobile communication (GSM) are reviewed. This section includes a detailed description of location and handoff management. This is important, as it describes how mobile units are located, what happens when they move from cell to cell, and the kind of communication that takes place.
The next part of the book reviews conventional database concepts. Transactions, serialization, and concurrency control are covered. The author then explains how mobility affects data management. Some of the issues introduced include location-dependent queries; the difficulty of enforcing atomicity, consistency, isolation, and durability (ACID) properties; and the effects of intermittent connectivity on transactions.
The last part of the book reviews various research directions for addressing the issues raised earlier. The research areas are grouped into a few major categories: transaction models geared to mobile and intermittently connected parties, alternate consistency and concurrency models, and transaction recovery issues in a mobile environment. The book ends with a chapter on wireless information broadcast.
The reader should understand that the book covers areas of research, as opposed to how things are currently done. I think the author succeeds in giving the reader a good sampling of the areas of research in mobile database systems. One can clearly see that people are revisiting every aspect of conventional data management in a mobile context. It is helpful to see this research gathered together in one book, because it allows one to see the different dimensions that are being explored.
The book does not describe how databases are used in real systems. In particular, the author does not explain how current cellular systems solve the data management problems that are raised in the book. Consequently, it is hard to evaluate the potential advantages of these new approaches. There is likewise a lack of data on whether these approaches are actually usable in practice. For example, we do not know how well any of the proposed transaction recovery models will work, or the performance implications of the new transaction models. Hopefully, the viability of these approaches will become clearer as time passes.