end to end argument

The main idea of this paper to provide a design view for function
placement in
different modules of a communication system. The paper brings out
point(later
referred to as guideline instead of being called as rule)
end to end argument - placement of a particular function
(for ex. which layer or module) should take into account knowlege of
end points and applications running at the end points. The authors first
cites
with careful filetransfer application which brings out the need for end to
end
argument, checksum at the endpoints and retry commit plan, to ensure
proper
working of the protocol. Then he continues to conclude that effort to
improve network
reliability at lower levels can have signficant improvement in performance
but that
itself can't guarentee perfect reliability rather it should consider
end to end perspective as well. He further cites different examples such
as
delay gurarentess, secure data transmission, to support his point.
Finally, he
concludes with the need to identify the end points in application and
subtlities
involved in identifying the same.

The papers is very well-written in the sense that author initially brings
out a
argument which he further supports by providing different examples and
finally
refines the arguement as a guideline instead of being thought of as a
rule.
Examples are very illustrative and provide insight into argument.

The author could have strenghened his arugment by doing measurement
studies on
communication systems that take into account end to end arguments vs those
that
don't. In conclusion, paper talks about the fact end to end arugment can
be
used to define different functions in different layers but author doen't
provide detailed explanations of how it can be acheivied in his paper.
Received on Thu Sep 14 2006 - 10:38:08 EDT