On 2009-08-06 05:34, Christopher Morrow wrote:
...
>> 2) Removing other "gems" (or clarifying them) like the second sentence in
>> the following:
>> ---cut here---
>> IPv6 nodes MUST NOT assume any mathematical or other properties of the Flow
>> Label
>> values assigned by source nodes. Router performance SHOULD NOT be dependent
>> on the
>> distribution of the Flow Label values. Especially, the Flow Label bits alone
>> make
>> poor material for a hash key.
>> ---cut here---
>
> 'flow label bits alone make a poor material for a hash key'... isn't
> this the reverse of saying that we'll (operators) require vendors to
> use flow-label for hashing on ECMP/LAG? If so, then... I don't think
> flow-label's going to cut it.
Please note the word "alone" in the above extract from RFC3697. That RFC
specifically does not forbid flows from different sources that happen
to go through the same router having the same flow label. That
would have required some massive unscaleable flow-label-distribution
protocol. (MPLS avoids this problem by having the downstream switch tell
the upstream switch what label to use.) So even if the flow labels are
pseudo-random, it's essential to include other material such as the
addresses in the hash. But clearly a pseudo-random flow label will
always improve the distribution of the resulting hash. Just don't use
it *alone*.
Our hidden fear when drafting those words was that lazy implementors
would allocate flow labels starting at 1 and counting up. That would
guarantee collisions in flow label space, strengthening the "poorness"
of the material. And it would also mess up any router whose performance
relied on the absence of flow label collisions.
I hope this clarifies, as requested. There's no reason to remove these
words, since they're true. The WG consensus at the time was to trim RFC3697
down to the normative text, so a lot of rationale was removed.
Brian
--------------------------------------------------------------------
IETF IPv6 working group mailing list
[email protected]
Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
