Hi Wendy, Thanks a lot for your always great insight. Your example is nice! I agree with you that supporting multiple topologies has many use cases, and feel that it is a key issue that ALTO will need to systematically support. For the example you give, will a single server, which provides multiple topologies, be easier to manage, than 3 servers? Push your example to the limit. Consider the deployment of one topology for each US state. Then we will have a lot more topologies. I am not sure if we want to say that IRD is dynamically customized.
Richard On Fri, Jul 12, 2013 at 1:21 PM, Wendy Roome <[email protected]>wrote: > Here's another take on that -- and a reason why an ALTO service provider > would need multiple Network Maps. > > I've long felt that a Network Map would have a varied level of detail: > detailed fine-grained PIDs near the client, coarser-grained as you get > farther away. For example, an ISP that covers USA might have different > network maps for eastern & western US clients. The Eastern one would have > 100 PIDs for the east, and 10 PIDs for the west. The western map would > reverse that. > > The ISP would run two different ALTO servers, one with the eastern map and > one with the western map, and would tune ALTO discovery to return the > server for the client's region. The result is that the discovered ALTO > server has only one map, but different customers discover different > servers. > > An advanced client that needs both network maps would have to "discover" > both ALTO servers via some other mechanism. > > - Wendy Roome > > On 07/12/2013 11:32, "[email protected]" <[email protected]> > wrote: > > >Having two separate servers would require two separate entry-points into > >the alto server discovery. Then the client has to pick a server at > >discovery time. This implies that ALTO server discovery has to find > >multiple servers for a client, which I'm not sure it is prepared to handle > >(and it was hard enough as it was). > > > > > _______________________________________________ > alto mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/alto >
_______________________________________________ alto mailing list [email protected] https://www.ietf.org/mailman/listinfo/alto
