Margaret Wasserman wrote: > The flow label does not, by itself, provide any useful information > that a router can use to classify a flow and/or optimize packet > handling. In fact, without knowledge of a specific signalling > mechanism or flow-establishment mechanism, the router can't > use the flow label for anything at all. For example, a router > cannot use a flow label for QoS queuing or to optimize hop-by-hop > header processing, unless the router is aware of the signalling/ > flow-establishment mechanism in use, since it will not > know the flow lifetime.
It is valid to say that a flow label can not be used in isolation, but I believe you are locking it too tightly to current implementations of QoS handling. If we consider that the flow label could be used as a modifier to a DSCP, flow lifetime is irrelevant, or more properly 'this packet'. One conceptual use of a DSCP/FL pair would be to manage congestion on a DSCP queue either dropping all packets of a single flow or ensuring the drops are spread across random flows. The choice would depend on the intended service level of the queue. > "Hosts or routers that do not support the functions of the Flow Label > field are required to set the field to zero when originating > a packet, > pass the field on unchanged when forwarding a packet, and ignore the > field when receiving a packet." [from RFC 2460]. I think the group is converging on modifing that text to: Hosts or routers that do not support the functions of the Flow Label field are required to set the field to zero when originating a packet, and ignore the field when receiving a packet. All routers are required to pass the field on unchanged when forwarding a packet. Tony -------------------------------------------------------------------- IETF IPng Working Group Mailing List IPng Home Page: http://playground.sun.com/ipng FTP archive: ftp://playground.sun.com/pub/ipng Direct all administrative requests to [EMAIL PROTECTED] --------------------------------------------------------------------
