Ad hoc networking sounds like something people do at cocktail parties, but actually it is a new branch of computer networking. Traditional networks and internetworks are based on the following model: there are a number of fixed switching machines, called routers, connected by wires or optical fibers. These form the backbone of the network. Each user machine is connected to a router through either a dial-up connection over a modem, a leased line, or a wireless (radio) link. When a process on one user machine wants to send a message to another user machine, it sends the message to the router to which it is connected. That router forwards the message to the next router along the path, and so on, until it reaches the destination. In this traditional model, used in all wide-area networks today, there is a very clear distinction between the user machines, which are owned and operated by the network customers, and the routers, which are owned by the network provider, often an Internet service provider. The routers and the connections between them are part of the fixed infrastructure, analogous to the telephone system, electricity grid, water pipes etc. found under every city.

Now imagine the following scenario: a major earthquake has destroyed the underground infrastructure. All the wires have been severed and most of the routers destroyed. The towers holding the base stations for cellular telephones have all toppled. Police cars, fire engines, ambulances, and other emergency vehicles rush to the scene. All of them are loaded with the laptop computers that the emergency workers have come to depend on for navigation, communication, and recording and for looking up information. None of the laptops, however, can communicate with their home bases anymore because the infrastructure has been destroyed.

This is where ad hoc networking comes in. Wouldn’t it would be wonderful if all the laptop computers could spontaneously form a network, setting up an appropriate topology so they could forward messages to each other, usually over several hops (because the radios in the laptops have limited range)? With a little bit of luck, the laptops closest to the edge of the pack would be within range of some fixed base station that was still functioning. Alternatively, police cars could drive to a series of intermediate points between the disaster area and the nearest functioning base station and stay there, forming a kind of multihop laptop link with the outside world. To make this work quickly, however, the laptops should be able to organize the network on the spot, with each one being both a user machine and a router.

Having a bunch of laptop computers organize themselves into a coherent network without human intervention is not trivial. It requires complex and highly robust algorithms. This book is all about those algorithms. After an introductory chapter, the rest of the chapters except the last one are about various algorithms of this type, explaining each one in detail and telling what its advantages and disadvantages are.

The last chapter provides a summary, and looks to possible future research, such as how big these ad hoc networks can get, what kind of quality of service they can offer, what happens when a client cannot locate a server, how these networks can deal with one or more malicious laptops in the network, and so on.

The book is highly technical, and different chapters are written by different authors, in varying styles. Few readers will be interested in so much detail about so specialized a topic, even with the understanding that it also applies to tanks on a battlefield, a fleet of ships at sea, and business people meeting out in the countryside. But if your job requires designing this kind of ad hoc network, this book is certainly worth reading.