[BitTorrent] An alternative "super-seed" mode

Olaf van der Spek OvdSpek at LIACS.NL
Sat Jan 29 18:29:08 EST 2005


Nick Johnson wrote:
> Jesus Cea wrote:
> 
> 
>>The current "super-seed" proposal is described in 
>>http://wiki.theory.org/BitTorrentSpecification and 
>>http://home.elp.rr.com/tur/superseed.txt.
>>
>>First, suppose that the seed is bandwidth "starved".
>>
>>I wish to discuss an alternative implementation, more simple:
>>
>>A "new age" Super-Seed node would behave like a usual seed, with a 
>>simple modification:
>>
>>When you are going to send a new piece, choose the peer that LESS pieces 
>>had received in the last X minutes (X=5, for example). Not from you, but 
>> from other peers.
>>
>>The logic here is that the peers receiving more pieces from others, are 
>>the ones sending less data to the swarm.
>> 
>>
> 
> So you're favoring peers that are parsimonious and don't contribute much 
> to the swarm? Given this strategy, the best strategy for a peer in a 
> super-seeder-heavy torrent would be to upload nothing, and hence get 
> priority from all the peers over those that are actually uploading 
> something.
> 
> It would seem to me to make far more sense for seeds to prioritise 
> sending chunks to nearly-finished peers first, to raise the number of 
> seeds as quickly as possible.

So that those new seeds can leave ASAP?
If you're using such a strategy, it's better to send pieces to peers 
that have almost nothing.


 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/BitTorrent/

<*> To unsubscribe from this group, send an email to:
    BitTorrent-unsubscribe at yahoogroups.com

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 





More information about the BitTorrent mailing list