On Mon, Sep 26, 2011 at 7:39 AM, Bill Roome <[email protected]> wrote:
> I have to agree with Sabastian. I think ALTO servers will be run by ISPs,
> and will have high detail and accuracy for the endpoints the ISP runs. But
> the detail and accuracy will drop as you get farther away from the ISP's
> territory.
>
> Example: An ALTO server run by a US-based ISP might divide North America
> into 100's of PIDs, and provide very accurate cost data between them. But
> that ALTO might have only 2 PIDs for all of Japan. The server might give
> good cost data for North America to those two Japanese PIDs, but it's cost
> data between those two Japanese PIDs would be very flakey.
>
> An ALTO server run by a Japan-based ISP would be just the opposite: lots
> of PIDs for Japan, but only a few for North America.
>

Sure - and that's what I meant when I said "depending on what you mean
by 'a high level of detail'".  The RFC5632 shows one example of a
"high level of detail" within
a single ISP, and for that trial, much less granular costs were given
to/from peering ISPs.  The ISP maps were not consulted when
determining costs between a pair of endpoints that were completely
outside of the ISP.  My only point was that the full map can still be
useful.  I'm not opposed to giving the ALTO Server a way of saying "I
don't know", which as Sebastian pointed out, basically makes that a
sparse matrix.  However, that's a separate thread...

Rich

>        - Bill Roome
>
>
> On 09/24/2011 14:03, "Richard Alimi" <[email protected]> wrote:
>
>>>
>>>But I doubt
>>>that there would be a single ALTO server that can deliver this matrix
>>>with a high level of detail.
>>
>>Why not? I think I would disagree with this, depending on what you
>>mean by "a high level of detail".  An ALTO Server does not need to
>>reply based on information from within other ISPs.  The semantics of
>>ALTO maps and information in general is that it is from the ALTO
>>service provider's perspective.
>
>
>
_______________________________________________
alto mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/alto

Reply via email to