[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [lisp] Confirming consensus behind echo noncing



Speaking very much as an individual.  I'm not objecting to echo
nonces.  However, I personally think we'll find they are not very
useful.

Echo-noncing is useful because it can avoid explicit control-plane probing. And we envision people will use active-active multi-homing so traffic will be flowing in both directions between a locator pair.

In particular,  they are not useful in any of the following situations

* square routing
* triangle routing (a sends to b, c sends to a)

The top 2 cases are one in the same.

* cannot detect a full path failure: in order to conclude you cannot reach someone you need to get packets from them

Echo-noncing is a unilateral protocol. So the other side will detect its forward path is down. You don't have to detect your return path is down.

I think that triangle and square routing will be very common unless we
take active steps to avoid them.  It seems likely that in any
situation where you have multiple rlocs of the same priority you'll
likely run into that case if you have a small number of flows.

You run into the case if one side is doing active-active and the other side is doing active-backup. But that means that echo-noncing can be used on each active side of each site.

Dino

Long term, especially when we take security considerations into
account, I think we'll end up with required control plane probing of
locators with possible optimizations through the data plane.  In that
environment, I think echo nonces will serve no purpose.  However this
is just my opinion.

I can't reason about or think about the performance implications until
I understand the deployment model of LISP.  In particular, the
performance concerns that matter for probing on CPEs seem very
different than say XTRs at an Amazon data center.

Regardless of the above, I think getting data on echo nonces can do no harm.
_______________________________________________
lisp mailing list
lisp at ietf.org
https://www.ietf.org/mailman/listinfo/lisp


Note Well: Messages sent to this mailing list are the opinions of the senders and do not imply endorsement by the IETF.