Richard, Sebastian, Wendy, All, Thank you very much for your comments.
I liked very much that Richard brought the issue of partitioned networks, which I was not thinking of before. The solutions for various sets may be totally different and we may have to discuss it thoroughly.
W dniu 2015-03-25 o 10:31, Sebastian Kiesel pisze:
2.1 and 2.2 are similar because both would require some kind of mesh between ALTO servers for exchanging authority information.
What I'm trying to solve though inter-ALTO is, among others, the issue that the current version of ALTO protocol can provide only a cost of uploading data, not downloading them. A local ALTO server (local ~= in the same ISP as the client) can easily gather all needed information (i.e., BGP upstream paths, IGP paths, link charging policies) to compare endpoints wrt. the cost of uploading data to them. Thus, a client can easily send a query to the server and get maps and/or a rating. Contrary, if a client queries ALTO servers in remote ISPs, it will get routing costs or ordinal ratings that it cannot compare, as each ISP applies its own metrics. Even if a local ALTO server ask remote ones for rating of its clients, it can do nothing with the gathered information. This issue is significant for both old (P2P) and recent (VPN, CDNs) ALTO use-cases. Using Sebastian's notation (thanks!), I was thinking of something like 2.1 + common method for exchanging an extract of network topologi information between ALTO servers.
Best, Piotr 'GhosT' Wydrych -- Piotr 'GhosT' Wydrych .. xmpp:wydrych//agh.edu.pl .. http://wydrych.net/ A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? _______________________________________________ alto mailing list [email protected] https://www.ietf.org/mailman/listinfo/alto
