On Mon, Sep 26, 2011 at 9:54 AM, Sebastian Kiesel <[email protected]> wrote: >> I agree that the quality of information (and whether >> the ALTO Server provides certain cost types for addresses outside of >> its administrative domain) is going to vary depending on many factors. > > > Let me rephrase my original point: > > If an ALTO client in an application running at the resource consumer > asks its ISP's ALTO server for guidance related to its own topological > location, the answer will be a (Mx1)-matrix (Vector). > > A tracker, in contrast, would be interested in the (NxN)-matrix. > > I'd say that usually M can be much higher than N, with high-quality data.
I'm curious - is the assumption that the tracker wouldn't be interested in high quality data? I'll note that I do agree with making the MxM matrix sparse (and avoiding the communication overhead of the entries where the ISP may have no idea). Thanks, Rich > > > > > Of course, the gut feeling "M >> N" does not allow any direct > conclusion whether tranmission of a (NxN) could become a problem. > >> FWIW, I fully expect there to be ALTO Clients out there who will crawl >> all publicly-available ALTO Servers and do data mining on that >> aggregated set. iPlane is an example of gathering data on a global >> scale without ALTO, and I would imagine that ALTO would only serve as >> another helpful input. But that is a separate question from the size >> of a map of a single ALTO Server. > > ACK. > > -- Sebastian > _______________________________________________ alto mailing list [email protected] https://www.ietf.org/mailman/listinfo/alto
