From: Kevin Wampler (wampler@cs.washington.edu)
Date: Mon Oct 04 2004 - 01:08:04 PDT
David Clark's paper "The Design Philosophy of the DARPA Internet
Protocols" gives a high level overview of the priorities that gave rise to
the Internet's current structure. In particular it points out how the
primary golds of survivability, support for multiple types of
communication services, and accommodation for a variety of networks
influenced the Internet's design, sometimes at the cost of secondary goals
such as ease of host attachment and accountability of resource usage.
The paper excels at filling in many of the philosophies in the design on
the Internet and relating them to the design decisions which they resulted
in. These relationships are further emphasized by frequent contrasts
between the designs chosen and other possible designs. This high-level
overview, of course, is by its nature somewhat limited in its description
of the lower level details, and at times presupposes a certain level of
knowledge about the workings of the current Internet (this is however
probably a safe assumption for most readers of such a paper). The paper
goes though some effort to mention where certain design decisions should
have been made differently, but details as to the reasons why another
method would be superior are somewhat scant. This fault is primarily a
matter of taste, but one of the few improvements I might suggest was to
either focus more on the specifics of the ails and remedies of the current
design, or leave out otherwise tangential discussions such as that which
constitutes the final paragraph of section 11. Other than this, I was
quite satisfied with the half historical half technical overview of the
Internet's origin.
The relevance of such a paper, other than relieving curiosity about the
birth of what has today become such a influential technology, is to aid in
the design of future communication networks. In particular, improvements
upon a technology are best done not only with knowledge of its faults, but
also with a firm grasp of the design philosophies and decisions which led
to these faults in the first place. The future work which be suggested
most naturally my such a paper would be the design of the next generations
of communication networks with a design philosophy informed both the
original design philosophies of the Internet and by the experience of its
strengths and weaknesses that have become evident over the years.
This archive was generated by hypermail 2.1.6 : Mon Oct 04 2004 - 01:08:04 PDT