From: Masaharu Kobashi (mkbsh@cs.washington.edu)
Date: Sun Oct 03 2004 - 12:50:46 PDT
1. Main result of the paper
It describes the Internet protocols and architecture in the historical
perspective explaining the motivations and evolution which led them to
their current state. It also sheds light on desirable future
improvements on the Internet design.
2. Strengths in this paper
First, unlike most other literature, it provides a historical
perspective and reasoning on why and how they have reached their
current state.
Understanding the historical development and process of evolution
gives us valuable insight for analyzing problems and forecasting future
needs.
Second, it provides very practical view of various problems such as
pointing out problems in the network design aids (page 111), most of
which, the author claims, lack the guidance on performance issues which
is vital in real implementation. Throughout the paper the same spirit
of practicality is alive.
Third, it gives us clear and detailed information on the motivation
behind the design of the Internet which is rarely available in most
other literature. For example, it explains in detail why TCP's flow
control is byte number based rather than packet number based.
Forth, it clearly shows the order of priority among the goals of the
Internet design. This information is rarely available and it alone can
explain some decisions which involve trade off relations among the
goals.
3. Limitations/shortcomings and suggested improvements
Overall it is well written and very informative. But there is still
room for improvement.
First, there is no discussion on the impact of the extremely fast
technology development of hardware and the reduction of hardware costs.
These factors can have substantial impact on the evolution of the
Internet in terms of design and demand/availability of types of
services.
Second, there is no discussion on the impact of changing pattern of
services and changing types of user groups on the performance of
the Internet. This discussion is important because it is related to
the Internet's core capability to accommodate wide variety of services
and networks and also because there has been constant changes in them.
Third, it would have been interesting if the author described what
would have been the design of the Internet if it were created from the
beginning for civil/commercial use instead of military objectives.
At least he should have pointed out which goals would have been removed
or added or modified if the Internet was designed for civil/commercial
use.
4. Relevance today and future
The paper has great relevance to both today's and future work on the
Internet since historical review can help us avoid future mistakes
and forecast potential dangers and opportunities.
The relevance in this respect is clear as the author explicitly states
what are the necessary improvements on the Internet. He claims they are
in the areas of resource management, accounting, and operations of
regions with separate administrations.
This archive was generated by hypermail 2.1.6 : Sun Oct 03 2004 - 12:50:47 PDT