| In the span of a quarter-century, network routing in communication networks has evolved tremendously. Just a quarter-century ago, the public switched telephone network (PSTN) was running hierarchical routing, ARPANET routing was operational, and the telecommunication infrastructure had fixed static transport routes. In the 1980s, we saw the first tremendous growth in routing: Internet routing was deployed under the TCP/IP stack starting, first with the RIP protocol; the telephone network started deploying dynamic call routing schemes; and the telecommunication transport network deployed SONET transport mechanisms, which could reroute in a ring topology in 40 millisec in the event of a failure. In the past fifteen years, we have seen the need for policy routing because of multiprovider settings, and the need to develop fast lookup algorithms for packet processing that enables efficient routing. We have also seen interdependency between addressing and routing as first addressed through classless interdomain routing (CIDR) and more recently, because of number portability in the PSTN.More importantly, we saw how the way an addressing scheme is deployed can impact routing and lookup algorithms.
Network routing can be broadly divided into three basic fundamental categories: packet routing, circuit-switched routing, and transport routing; certainly, a combination is possible. The evolution over the past quarter-century has brought to the foreground the need to understand and examine where and how different dimensions of routing, from algorithms to protocols to architectures, can differ for different types of networks, and where they intersect. Certainly, the goal is to learn from our past experiences and prepare ourselves for next generation networks and routing.
While numerous papers have been written on the subject of network routing, and several books are now available on routing for specific networks, the field still lacks a comprehensive or systematic guide that encompasses various routing paradigms. Second, even in discussions of a single routing type (for example, either the Internet or PSTN), the focus often appears to be either on protocols or algorithms without tying them together with analysis and implementation; or, the work delves more into router command-line for router configuration; or, being informationalwithout explaining thewhys. Furthermore, how the addressing mechanism can affect routing decisions is yet another important topic that is rarely discussed. For efficient routing, how routers are architectured—and why—is yet another mystery. Finally, the relation between traffic engineering and efficient routing is also another topic. In the end, one needs to be somewhat of an “expert” in different routing paradigms to get a well-rounded view. |
