Review - Where Wizards Stay Up Late: The Origins Of The Internet by K. Hafner and M. Lyon

For those unsatisfied with the customary paragraph on the history of Internet in computer networks books, this is just the text you may have been looking for. The book charts the journey from the early conceptions of a packet switching network up to the point where this network was poised to explode onto the world stage, becoming the Internet we know of today. Along the way the reader meets several characters that are otherwise familiar from the many papers and RFC's that they authored - Jon Postel, Wes Clark, Bob Kahn, Steve Crocker, Vint Cerf - along with other less familiar ones that shaped the early internet. (Incidentally, RFCs started as meeting notes kept by Steve Crocker - with the first one being written in a bathroom to avoid disturbing roommates - who choose the name Request for comments (RFC) to avoid sounding too declarative).

The book is filled with interesting incidents, starting with the rejection of the packet switching idea by ATT, which had too much invested in telephone virtual circuits to consider the idea on its merit. Such a network then gets developed by BBN (Bolt Beranek and Newman) , a small company which to everyone's surprise wins the ARPA contract for this. Ironically, later its BBN turn to bury its head in the sand when it fails to have the vision to capitalize on its success by refusing to go into the router business.

Technical details and controversies are presented in a manner that make will make them interesting to the lay readers as well as engineers. To recount just a few.

Initially congestion control and reliability measures get built into the first routing devices - the IMPs or Interface message processors as they were called. Only later, when different networks - the wired ARPANET, satellite based SATNET and radio waves based Packet Radio network, each with different reliability characteristics need to be linked together does the idea of moving these functions to the hosts gain currency. This then gets refined into the transmission control protocol TCP.

We see Bob Kahn arguing for correcting the inherent deficiencies in the original algorithms and design, only to be ignored by the engineers who are too busy building the thing to worry about any 'hypothetical' network scenarios. Kahn of course was right and able to demonstrate this on the initial versions of the network. Effort to iron out these problems is then undertaken.

The reader is made witness to the OSI vs TCP/IP protocol wars. TCP/IP is considered everything from experimental to being suitable only for toy networks. TCP/IP however has the advantage of already working while OSI has only the pedigree of its advocates to show for. TCP/IP, as we know wins, giving credence to the mantra that best solutions are discovered not decreed.

To sum up, this book is a highly readable account of the journey of the Internet's beginnings . It educates even as it entertains. Strongly recommended for anyone interested in this subject.