(no subject)

Paper Review: The Design Philosophy of the DARPA Internet Protocols

This paper examines the Internet protocol suite, TCP/IP, developed by DARPA.
The paper extends previous works, which focused on the operation of the
protocols, by providing the motivation behind the design aspects of the suite.

The first goal of the architecture was to be able to interconnect existing
networks to create a single internetwork. Packet switching, best-suited for
early applications, was chosen for multiplexing. The networks were
interconnected using store-and-forward packet switches based on the original
ARPANET project. A secondary set of goals, in order of importance, was also
presented. The paper notes that the project was developed in a military context
rather than commercial, explaining the ordering of the goals.

Survivability of the Internet was provided by storing state information in the
end-systems thereby protecting it (fate-sharing). A variety of services can be
supported through multiple transport layer protocols (TCP & UDP). Variation in
network types is made possible by making very few, simple, assumptions about the
underlying architecture.

The paper furthers the research by examining those features of the protocol
suite which make it less than perfect, yet explains why certain decisions were
made out of necessity at the time of creation.

This paper’s contribution to network research is that it describes, with great
detail, how each feature of the protocol suite came to be. Since the Internet
is perhaps the world’s most widely used system, systems designers can look to
this paper when making their own decisions. The paper supports all its
explanations with concrete evidence and is very thorough in all its explanations.

-- Nadeem Abji
Received on Thu Sep 14 2006 - 03:57:15 EDT