Re: BitTorrent swarms have a deadly bite on broadband nets

["Steven M. Bellovin" <smb () cs columbia edu>]

On Sun, 21 Oct 2007 13:03:11 -0400 (EDT)
Sean Donelan <sean () donelan com> wrote:

> http://www.multichannel.com/article/CA6332098.html
>
>    The short answer: Badly. Based on the research, conducted by Terry
> Shaw, of CableLabs, and Jim Martin, a computer science professor at
> Clemson University, it only takes about 10 BitTorrent users bartering
> files on a node (of around 500) to double the delays experienced by
> everybody else. Especially if everybody else is using "normal
> priority" services, like e-mail or Web surfing, which is what tech
> people tend to call "best-effort" traffic.
>
> Adding more network bandwidth doesn't improve the network experience
> of other network users, it just increases the consumption by P2P
> users. That's why you are seeing many universities and enterprises
> spending money on traffic shaping equipment instead of more network
> bandwidth.
This result is unsurprising and not controversial.  TCP achieves
fairness *among flows* because virtually all clients back off in
response to packet drops.  BitTorrent, though, uses many flows per
request; furthermore, since its flows are much longer-lived than web or
email, the latter never achieve their full speed even on a per-flow
basis, given TCP's slow-start.  The result is fair sharing among
BitTorrent flows, which can only achieve fairness even among BitTorrent
users if they all use the same number of flows per request and have an
even distribution of content that is being uploaded.

It's always good to measure, but the result here is quite intuitive.
It also supports the notion that some form of traffic engineering is
necessary.  The particular point at issue in the current Comcast
situation is not that they do traffic engineering but how they do it.


                --Steve Bellovin, http://www.cs.columbia.edu/~smb