Re: [p2pi] Information in an ALTO protocol

On Sep 15, 2008, at 8:06 AM, Nicholas Weaver wrote:

>
> On Sep 15, 2008, at 7:54 AM, Laird Popkin wrote:
>
>> The cost of bandwidth testing isn't terrible, but the accuracy of
>> measurements taken on 1K transfers is very low - you're really
>> measuring latency and buffering, not sustained data transfer rate.
>> Of course, there might be some relationship between them, but (at
>> least in our measurements) it took much larger data transfers to get
>> even remotely stable results. In addition, measuring actual
>> throughput at a point in time can be affected by other traffic on
>> the PC (e.g. PC's do many things when they're starting up), other
>> nearby PC's (e.g. someone else in your neighborhood is busy right
>> now), transient interference on the internet, etc. So while it's not
>> unreasonable to do such testing, and we certainly do so, an ISP
>> provided provisioned capacity would be much more accurate (when it's
>> available).
>
> The point of the 1K back-to-back packets is NOT that you are testing
> the transfer time for the 1K packets, but that interarrival time of
> the packets can be used as an uncongested-link bandwidth estimate, as
> long as the queue at the point of congestion uses FIFO delivery of
> packets.
>
> Since the question you are asking the user is "whats the UNCONGESTED"
> behavior as a starting point (to set an initial threshold/ceiling
> somewhere below this point), the behavior of back-to-back packets
> should get you in the right ballpark without having to ask the user.

Nick,

I think you're talking about a theoretical "should" and Laird is  
talking about what he's observed in practice.

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