bittorrent port range?

[Andrew Kroeger]

Fulko.Hew at sita.aero wrote:
> 
> fedora-test-list-bounces at redhat.com wrote on 09/14/2006 03:00:10 PM:
> 
>> G.Wolfe Woodbury wrote:
>>> I could have sworn that I saw a notice of what range of ports
>>> bittorrent wants opened
>>> to optimize performance, but reviewing the archive I can't find
>>> the numbers right now.
>>>
>>> Could someone give me a clue of what ports to forward to the
>>> torrent box from my router?
>>>
>> bt can use 6881 - 6999
> 
> Since I've had some issues (in the past) regarding running BT at a
> respectable data rate even though I have the ports open and the
> firewall forwarding...
> 
> I'm wondering if some ISPs aren't blocking BT ports now
> to prevent RIAA and MPAA 'theft'?
> 
> Any comments?
> 

Yep, that's my situation.  I don't use BT that often, so I first noticed 
the "problem" when I tried to download the FC6t2 ISO.  I initially 
thought I had screwed up my firewall configuration and almost drove 
myself nuts trying to debug it.

What I finally discovered was that netstat -n showed many ESTABLISHED BT 
connections (ports around 6881), but my BT client only showed 1 active 
peer.  Looking into that active peer showed that it was running on a 
non-standard BT port -- don't remember exactly what, but it was +9000. 
I checked with my ISP and they confirmed they are blocking BT and other 
P2P protocols.

I changed by BT client to listen on a non-standard port, and still saw 
the same behavior -- only 1 active peer, using a non-standard port.

Given that a lot ISP's are probably like mine and only use simple port 
range blocking (as opposed to analyzing the packet data) to clobber BT 
and the other P2P technologies, maybe it's time for everyone to change 
their clients to use non-standard ports.  The trackers take care of the 
port changes, so this problem is already solved in the BT protocol.  The 
hardcore P2P users are already doing this -- some are even using 
encrypting BT clients to avoid attempts at data analysis -- and those 
are probably the users most likely to draw an RIAA/MPAA nastygram.

I see no problems in using this technique to avoid the blocking.  It 
gets releases out to a wider audience (as opposed to those who have BT 
blocked and simply give up), which gets more hands/eyes on using/testing 
the releases that the ISP's will eventually use to provide us with their 
"services" (or blocking thereof).

Andrew