Thanks to Ben and Rich for the respond. I agree with your answers. BR, -Haibin
> -----Original Message----- > From: [email protected] [mailto:[email protected]] On Behalf Of > Richard Alimi > Sent: Monday, December 26, 2011 10:55 PM > To: Ben Niven-Jenkins > Cc: Songhaibin; [email protected] > Subject: Re: [alto] One small comment to ALTO protocol > > On Mon, Dec 26, 2011 at 6:50 AM, Ben Niven-Jenkins > <[email protected]> wrote: > > Haibin, > > > > On 26 Dec 2011, at 09:42, Songhaibin wrote: > > > >> The current ALTO protocol document describes the dependency of the cost > map and network map, and only network map version tag is used for consistence > check. But the cost map might be changed without the change of network map, I > think we should allow the ALTO client to only update its cost map if the > network > map has not been changed. > > > > This seems an unnecessary restriction as if the Network Map changes in many > cases the resulting Cost Map will have to change and in addition there will be > cases where the Cost map changes when the Network Map stays the same. > > > >> Shall we add a version tag for the cost map? > > > > My preference would be that ALTO Servers & Clients make use of HTTP Etags so > that clients can do IfNoneMatch type operations to check if the Cost Map has > changed without having to first obtain the entire unchanged map again and > check > for a costmap-version field embedded in the map itself. > > > > Agreed. We only needed a version tag for consistency between network > map and cost map because they are two separate HTTP resources and we > need a way to cross-reference them. Asking for updates to single > resource can be done using the standard tools in HTTP. > > Rich > > > Ben > > > > _______________________________________________ > > alto mailing list > > [email protected] > > https://www.ietf.org/mailman/listinfo/alto _______________________________________________ alto mailing list [email protected] https://www.ietf.org/mailman/listinfo/alto
