Wendy, Thanks for clarifying the alternatives. I also thought that we agreed on #2, and the sentence at 11.5.1.5 is a left over. I think a clean, simple change is #2: "ECS MUST NOT use a Network or Cost Map. Hence, the ECS cost is the cost from the source to destination endpoint. A future extension may allow ECS to state that it “uses” a Network Map. The extension then will need to define the semantics.” Does this work for you?
Richard On Feb 25, 2014, at 3:25 PM, Wendy Roome <[email protected]> wrote: > Richard, > > I don’t like that language — it’s too wishy-washy. If an ECS can “use” a > Network Map, we need to spell out what that means. > > So I support either of two alternatives. > > (1) Say that an ECS MAY “use” a Network Map, and spell out what that means. > E..g., if an ECS “uses” a Network Map, then that ECS is equivalent to a > Filtered Cost Map resource on that Network Map, using the PIDs of the source > and destination endpoint addresses. This means the client can assume the ECS > will return the same cost for any endpoint in the source PID to any endpoint > in the destination PID (modulo cost updates). > > If an ECS does not “use” a Network Map, then the client can make no > assumptions about how the ECS costs relate to the costs for any PIDs or > endpoints. > > (2) Say that ECS MUST NOT “use” a Network or Cost Map. The ECS cost is the > cost from the source to destination endpoint; the client cannot draw any > conclusions about costs for PIDs or other endpoints. > > BTW, I thought we’d agreed on #2, until a colleague pointed out that 11.5.1.5 > says an ECS can use a map. > > - Wendy Roome > > From: "Y. Richard Yang" <[email protected]> > Date: Tue, February 25, 2014 at 10:52 > To: Wendy Roome <[email protected]> > Cc: IETF ALTO <[email protected]> > Subject: Re: [alto] Ambiguity in ALTO draft 25 > > Hi Wendy, > > Good discussion. I believe that the previous conclusion is that ECS does not > reveal any dependency. Hence, a simple change will be the last sentence: > > "...However, to preserve flexibility, there is no need for the ECS resource > to declare the network map and/or cost map on which it depends." > > How does this sound? > > Richard >
_______________________________________________ alto mailing list [email protected] https://www.ietf.org/mailman/listinfo/alto
