young Fedora torrents spread data too slowly

[Callum Lerwick]

On Sun, 2008-10-12 at 11:26 +0200, Denis Leroy wrote:
> Mike McGrath wrote:
> > I'll have a talk with Jesse.  For the normal releases we typically have
> > multiple seeders from the start.  For these sorts of snapshots I suspect
> > some steps can be taken to ensure better usability here.  I talked with
> > Seth a bit about this too (he's our current torrent wrangler) and we
> > generally think this was just an issue with two many chunks and not enough
> > people for chunks from the beginning.  Which is self correcting but A)
> > is annoying and B) is preventable in the future.  We just have to figure
> > out the procedure on it.
> 
> I'm not convinced by this explanation. If you know a little bit about 
> the bittorrent protocol, the scenario of a fast uploader and about 40 
> downloaders is literally a textbook example where bitorrent behaves at 
> its best. Yet the algorithm broke down fairly quickly, as though the 
> server pushed the chunks in linear order. That or something on the 
> network got in the way.

Any followup on this? I doubt it's the protocol/algorithm. It doesn't
matter what order the seed "pushes" blocks, if a seed gives a block to
just one of a well connected set of "leeches", that block will quickly
spread to all the others, causing them to never request that block from
the seed. Bittorrent block transfers are initiated by request, not
"pushed" by the seed.

It's far more likely that the seed is throttled in some way, than if
every one of those downloading peers are not talking to each other...

I'd look for throttling somewhere near (or at) the seed. Some ISPs are
getting very aggressive about throttling P2P traffic. A major backbone
throttling bittorrent traffic is quite likely these days... :P
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part
URL: <http://listman.redhat.com/archives/fedora-devel-list/attachments/20081016/b8d7437a/attachment.sig>