From: Yuhan Cai (yuhancai@cs.washington.edu)
Date: Sun Oct 03 2004 - 10:39:13 PDT
Title: The Design Philosophy of the DARPA Internet Protocols
Author: David D. Clark
Reviewed by: Yuhan Cai
While most existing work is focused on the specifications of TCP/IP, it
has been less clear that what kind of design philosophy researchers have
in mind when they develop the internet suite. In this paper, Clark
presents the important goals of TCP/IP, the underlying reasoning behind
them, and the related key features of the suite.
One of the strengths of the paper is that, it not only identifies the
fundamental goal of TCP/IP, but also prioritizes other second level goals
with respect to different areas of applications. First of all,
multiplexing is claimed to be the top level goal and therefore packet
switching deserves the most important place in the internet architecture.
Secondly, the author outlines a list of secondary objectives and orders
them by their importance. The ordering is then justified according to
different domains. In this way, Clark has provided people a blueprint of
the internet and an outline of what they should do and what they should
not do in conducting their research. For example, if they were to develop
a brand new internet protocol, packet switching would be the last thing
they want to touch upon. One possible weakness of this paper is that, it
does not address the issues of security, which turns out to be one of the
most crucial problems in the internet age.
Although this paper was published over 15 years ago, it is still relevant
to the research communities nowadays in that it enables developers to have
a clear picture of the original motivations, governing principles and
necessary requirements of TCP/IP. One direction of future work it suggests
is to explore for better building blocks than datagrams. Another direction
is a complete review of the history of TCP itself.
This archive was generated by hypermail 2.1.6 : Sun Oct 03 2004 - 10:39:14 PDT