Re: [v6ops] Flow Label Load Balancing

[Fernando Gont <[e-mail]>]

On 24/11/20 16:28, Alexander Azimov wrote:
> Hi Fernando,
> 
> Stating that FL change during TCP session lifetime is a bug - is a bit 
> harsh.

Apologies if it came across like that -- the intent was to be blunt, 
rather than harsh.




> It is a fantastic idea to change the FL value if RTO or SYN_RTO happens 
> in a controlled environment.

As noted in my previous email, I'm not sure what you mean by RTO -- but 
if you really mean "retransmission timeout", doing that fails the 
definition of FL by RFC6437 and others. -- a retransmitted packet 
belongs to the ame flow, not to a different one.

That aside, please define "controlled environment". :-)



> These are very specific TCP timeouts, that provide enough guarantee that 
> there will be no out-of-order packets,

Strictly speaking, that's incorrect:: the RTO is an estimation, and may 
be fired incorrectly (although the whole point of measuring the RTT anc 
computing the RTO is for that case to be rather rare).


>  though your packets will reach 
> the destination even in case of an outage inside your network. Zero 
> influence on your services in case of the network outage - doesn't sound 
> like a bug for me.

It seems to me that this sort of behavior is the reason for which folks 
actually disable the use of the FL. (i.e., think what happens for load 
balancers).

Thanks,
-- 
Fernando Gont
e-mail: [e-mail] || [e-mail]
PGP Fingerprint: 7809 84F5 322E 45C7 F1C9 3945 96EE A9EF D076 FFF1