Re: [p2pi] Refining the ALTO problem statement

For what it's worth...

>>   It is reasonable to expect P2P applications to learn how to use more
>> than one remote peer at a time. Statistical information is never a
>> guarantee that you'll see exactly the predicted situation.
>
> True, but unless the information is accurate much more often than not, 
> why bother designing a mechanism to provide it? Apps will just ignore  it 
> then.

This was actually the stake through the heart of TRIGTRAN, if anyone 
remembers that conversation - TCPs are complicated enough already, so if you 
have to treat non-loss indications as advisory, and you can figure out what 
the right answer is in a small number of RTTs without using the other 
indications anyway, the thought was that there wasn't enough reason for 
implementers to find TRIGTRAN interesting.

> I question whether we actually can design an oracle that is right much 
> more often than wrong with regards to performance-related information. 
> Trying to answer this question as a research effort is very  interesting, 
> doing standardization around it is IMO premature.

You may very well be correct... :-)

Spencer 

_______________________________________________
p2pi mailing list
p2pi@ietf.org
https://www.ietf.org/mailman/listinfo/p2pi