From: Craig M Prince (cmprince@cs.washington.edu)
Date: Sun Oct 03 2004 - 23:32:47 PDT
Reading Review 10-04-2004
-------------------------
Craig Prince
David Clark's paper "The Design Philosophy of the Darpa Internet
Protocols" describes the design decisions that went into the development
of the IP and TCP protocols -- giving motivations for the protocol's
design decisions and explaining why alternatives were not chosen. The
major contribution of this paper is to give insight into the history of
the current internet protocols and to highlight the intentions/assumptions
of the designers of these protocols. This is accomplished by first listing
the (prioritiezed) goals of the designers and then explaining how the
protocols were built/revised to match these goals.
I found that this paper did a good job of accomplishing its task of
explaining the goals of TCP and IP and giving a good analysis of the
strengths and weaknesses of each of the design decisions -- starting with
the main goal and then addressing each of the secondary goals (in order of
importance). This organization was effective in showing how, by addressing
the high-priority goals first, it was increasingly difficult to satisfy
the remaining lower-priority goals. This organization allowed me to see
the various tradeoffs that are made when designing the protocols.
While many of the explanations given in the paper seem reasonable and
logical, I was concerned by the fact that the author does not give any
intuition as to how he knows that his explanations are the correct
explanations behind the goals and decisions he describes. The author does
mention that he was involved in the design of these protocols but doesn't
mention whether his explanations are first-hand or derived from the
experiences of others. I believe that this paper could have been improved
by offering more of the arguments against the design decisions made. The
paper gives good motivation for the design decisions, but there must also
have been reasonable arguments for alternative architectures.
Even today this paper is extremely relevant in that it highlights the
strengths of the current network protocols and suggests different
directions for research motivated by the demands and requirements of
networks today. The paper gives insight into the limitations of the
current internet protocols and lists several challenging problems for
these protocols (guarantees on bandwidth, mulit-cast, resource
accounting).
This archive was generated by hypermail 2.1.6 : Sun Oct 03 2004 - 23:32:48 PDT