(no subject)

Designing complex systems like a distributed computer system requires the
system designer to properly identify the boundaries of different functions
within the system and placing these functions at the right level within
system modules. In this paper the author present a design principle, an
end-to-end argument, that the author claims can help the designer to
places functions at appropriate modules within a system. According to
end-to-end argument that if a software requires some sort of function at
application level then there is no need to implement the same function at
lower level because it will not provide any extra benefit neither will
increase the performance. In order to support his view the author has
presented an example of data communication network and has shown that if
the application requires encryption and acknowledgements at the
application level then providing these functions at lower levels will
neither increase the performance nor reliability. Also the author has
discussed general example e.g. airline reservation system, telephone
exchange etc to show how using an end-to-end argument can help design an
efficient system but for applying end-to-end argument the designer should
carefully identify the end points of the application to which the argument
is to be applied.
Systems are becoming complex and at the same time the demands of the user
for performance, reliability, security etc are also increasing. So the
systems that are being build need to efficient as well as reliable but in
order to cope these requirements a system designer need to clearly
identify the system modules and their functions and the end-to-end
argument can provide guidance in this respect.
Received on Thu Sep 14 2006 - 09:06:12 EDT