On 5/2/13 2:24 PM, Fred Baker (fred) wrote:
On May 2, 2013, at 2:17 PM, Brian E Carpenter <[email protected]>
wrote:
That's why I think the way out is to use the wiggle room mentioned
above. I hope we can.
I'm afraid I don't see any wiggle room. Section 3 of RFC 6437 requires every
new flow - every new TCP session, in the most extreme reading of that - to have
a new flow label value. This proposal presumes that all of the flows subject to
the same security policy would be identified by the same flow label. By your
rubric, an operator who is not using the flow label for load balancing MUST NOT
use the flow label for a different purpose. Frankly, I wish you luck enforcing
the ruling. Operators have a funny habit of doing what they deem important. Our
job in the IETF is to help them be able to do that using software and hardware
from multiple interoperable sources.
I've been uncomfortable with the attempts to rehabilitate the flow label
since they seem to be inconsistent. I can't do something useful if
that's largely unusable due to host behavior, if I keep changing what
the hosts are doing with it.
If I can resuse it without the hosts changing their behavior, great
especially if it's within my own span of control. but then I don't
really need ietf consensus for that.
--------------------------------------------------------------------
IETF IPv6 working group mailing list
[email protected]
Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
--------------------------------------------------------------------
IETF IPv6 working group mailing list
[email protected]
Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
