[bittorrent] IO bound

Alan McGovern alan.mcgovern at gmail.com
Fri Apr 13 10:49:00 EDT 2007


Looks like we'll have to agree to disagree otherwise this could turn nasty
;)

IO problems are exactly why the SuggestPiece message was introduced. Hence
why i'm advocating that you use a library that supports the FastExtensions
to "fix" the problem. I would hope that every library claiming full fast
peer support would only do so if it handled all messages in a "sane" manner.

All you need to rely on is that a client claiming to support the Fast
extensions doesn't completely ignore messages defined in the extension. That
shouldn't be an unreasonable expectation.

Alan.

On 4/13/07, Adrian Ulrich <torrent at blinkenlights.ch> wrote:
>
>
> > Sorry, yes. I meant AllowedFast. And it does affect chunk selection ;)
>
> True, it does.
> Didn't think too much about it because in my client
> allowed-fast just creates an internal 'fake-bitfield' for this peer.
> The selection itself doesn't really change/know about this.
>
>
> > In other words, download the suggested piece unless you have a good
> reason
> > not to. Thats the way i've treated the message, and it makes the most
> sense
> > to me.
>
> ACK. But to me 'i do not care' is also a good reason ;-)
>
>
> So let's summarise:
>
> - Use fake-bitfields if you need to enforce selection
>    (Having IO-Problems would fit into this)
>
> - Use SuggestPiece if you trust your peers / do not
>    need to enforce piece selection.
>
>
>
> --
> RFC 1925:
>    (11) Every old idea will be proposed again with a different name and
>         a different presentation, regardless of whether it works.
>
> _______________________________________________
> BitTorrent mailing list
> BitTorrent at lists.ibiblio.org
> http://lists.ibiblio.org/mailman/listinfo/bittorrent
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.ibiblio.org/pipermail/bittorrent/attachments/20070413/07030707/attachment.html 


More information about the BitTorrent mailing list