[p2p-hackers] Using UDT for swarming

Serguei Osokine Serguei.Osokine at efi.com
Tue Jan 25 23:11:08 UTC 2005


On Tuesday, January 25, 2005 Justin Chapweske wrote:
> The "long fat" and "different RTT" are really two faces to the 
> same problem, so solving one should take care of the other. Both
> are caused by the fact that high RTT flows get pummeled by packet
> loss, while low RTT flows are able to bounce back quite quickly 
> after packet loss.

	Fair enough - thanks! (Sometimes "long fat" is understood as
"saturating the long fat pipe" without any special attention paid
to the fairness issues - for example, by UDP-bombing the link, the
other streams be damned. Just wanted to make sure that we're on
the same page here; sorry for insulting your intelligence :-)

	Best wishes -
	S.Osokine.
	25 Jan 2005.

-----Original Message-----
From: p2p-hackers-bounces at zgp.org [mailto:p2p-hackers-bounces at zgp.org]On
Behalf Of Justin Chapweske
Sent: Tuesday, January 25, 2005 2:36 PM
To: Peer-to-peer development.
Subject: RE: [p2p-hackers] Using UDT for swarming


On Tue, 2005-01-25 at 14:13 -0800, Serguei Osokine wrote:
> On Tuesday, January 25, 2005 Justin Chapweske wrote:
> > The FLUTE/ALC/LCT stack does indeed solve the long fat network 
> > problem and we have a number of customers deploying the solution 
> > for transfer of extremely large data sets (100 GB+) over both 
> > satellite and terrestrial networks.
> 
> 	So that would be solving only "long fat" or also "different RTT"
> problem? That is, do you handle simultaneous terrestrial (low RTT) and
> satellite (high RTT) streams without the latter getting lower bandwidth
> than it should due to the unfairness of competition with the former?

The "long fat" and "different RTT" are really two faces to the same
problem, so solving one should take care of the other.  Both are caused
by the fact that high RTT flows get pummeled by packet loss, while low
RTT flows are able to bounce back quite quickly after packet loss.

-Justin

_______________________________________________
p2p-hackers mailing list
p2p-hackers at zgp.org
http://zgp.org/mailman/listinfo/p2p-hackers
_______________________________________________
Here is a web page listing P2P Conferences:
http://www.neurogrid.net/twiki/bin/view/Main/PeerToPeerConferences



More information about the P2p-hackers mailing list