From: T Scott Saponas (ssaponas@cs.washington.edu)
Date: Sun Oct 03 2004 - 23:49:07 PDT
"The Design Philosophy of the DARPA Internet Protocols" presents one viewpoint as to the original thinking behind the design decisions of the DARPA Internet protocols. The main contribution of the paper is an examination of the relationship between the original goals of the protocols and their actual features. The impact of the ordering of the stated goals on the protocols is shown through describing features such as "fate-sharing" and how such a feature aids in meeting the first goal of survivability but makes other goals such as QOS and resource accounting more difficult. The paper also does a good job in linking the roots of some of the drawbacks to the protocols, such as the lack of ability to formalize performance constraints within the architecture, to the climate under which the protocols were designed. One interesting example of shedding light on the source of a design decision is originally TCP and IP had been a single protocol, but the need for more than one transport service forced them to be separated into two layers making only IP the center of the networking hourglass.
While the "TCP" section of the paper is interesting, it strays a little from the form of the rest of the paper. This section contains interesting discussion of some of the particulars of TCP and how it evolved, but it could be made more relevant to the rest of the paper if it was discussed more in terms of how those parts of TCP relate to the stated goals of the protocols.
This paper suggest future work in Internet protocols might consider something besides datagrams in meeting some of the rest of the goals not at the top of the list, like resource management and accountability, since the Internet is no longer just a US defense project.
T. Scott Saponas
This archive was generated by hypermail 2.1.6 : Sun Oct 03 2004 - 23:49:08 PDT