[e2e] variable bandwidth link
ddutta at ISI.EDU
Tue Apr 17 19:36:32 PDT 2001
In some cases, abstracting the network dynamics with the help of variable
bandwidth link may not be a bad proposition. Lets assume you have a
heterogenous network with varying link condition and you want to study the
effects of TCP. A crude abstraction could be to reduce it by one special
link that models the hetereogenous network.
> Of course, the practicality of this variable bandwidth model depends on
> the practicality of your assumption about hiding wireless errors
> from TCP, which may be quite practical in some situations (and not so in
> many others). I think it's a good approach though.
In a way, the above reasoning works. But sometimes, hiding lower layer
abstractions may not be the most important reason. One can even model the
effects of the motion of a mobile host as a variable bandwidth link from
the end2end perspective. I understand that this is not easy but one can
assume some standard models and see the effect of TCP on them.
More information about the end2end-interest