Margaret Wasserman wrote:

> As far as I can tell, this makes it impossible to use the flow label
> as current defined (without some out-of-band flow management system
> to supply knowledge of flow lifetimes) to identify a single flow.
>
> What am I missing?

As I said earlier, you are focused on using the FL in issolation, and
finding that doesn't make sense. In that context you are correct, but
your assumption that all administrative domains in the path need to know
the same context as all others seems to be leading to your confusion. If
an intermediate domain wants to use a static flow management system that
says for the DSCP queue XYZ, when congestion occurs drop all packets
with the same flow until congestion subsides, but for DSCP queue ABC
ensure that drops are spread evenly across all flows, that is as valid
as any signaled system that might be used at the edge. If a source
happens to reuse a FL within the window of congestion in the
intermediate domain, so be it. This is not a particularly likely
scenario since most implementations will use existing randomization
routines to acquire the value for the FL, so why are we still talking
about it?

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]
--------------------------------------------------------------------

Reply via email to