On Thu, Sep 14, 2000, Gerald Oskoboiny wrote:
> For a single large file, I don't really see how this would speed
> things up, unless the multiple connections are just getting
> around some kind of artificial limit per connection or something.
> (shouldn't a single download be enough to saturate whatever
> bandwidth you have available?)
On Thu, Sep 14, 2000, Joseph M. Reagle Jr. wrote:
> But then again, if this accelerator really does work that means the
> intermediary pipes can handle the bits (otherwise I wouldn't see a
> difference) AND the server can as well. So was the previous "inefficiency"
> arbitrary; or am I benefiting at the expense of my network peers?
This is actually a very interesting question: what is the bottleneck for
a TCP connection between two hosts?
I guess that the answer is simple: at some point along the route between
the source and the destination, there is a congestion.
Being very simplistic:
If we imagine that there are 100 TCP connections at this particular
point, and that this is the only problem along the way, the bandwidth
you get with this connection would be 1/100th of the available
bandwidth.
If all things are equal and you open 4 extra connections, you will get
5/104th of the bandwidth, i.e. that you will download your large file
4.8 times faster.
Again, this is very simplistic and I'm not even sure that it would work
that well, but I guess that it is the idea.
And if everybody starts using download accelerators, a user will end up
with 5/500th of the bandwidth, i.e. the same as the beginning, except
that there will be 5 times as many TCP control packets as there were
with only one connection, which means that this increased overhead will
make the user lose bandwidth used for the actual transfer.
--
Hugo Haas <
[email protected]> -
http://larve.net/people/hugo/
I would kill everyone in this room for a drop of sweet beer. -- Homer
J. Simpson