Hi Hannes,

> we have heard of many customer requirements requiring to to disjoint
> routing

True if you are talking end to end between PEs. I have heard of those
between 2004-2006 when MTR fashion was at the peak too :)

And I must have missed that SR claims to address those without all
nodes supporting SRH processing if you are talking PE-PE.

But there are cases where to the PE level is not required for number
of applications so those can be addressed without all nodes handling
SRHs.

Again we are not here to decide I think.

If customer takes a calculator and conclude that for his need it is
better to deploy MPLS data plane he will dop that no matter for SPRING
or FALL brings to the table.

If on the other hand he will see that just he can upgrade IOS on bunch
of nodes to address his needs he will do just that.

> with LSP hierarchy we can solve pretty much any scaling problem, right ?

Yep - never claimed we can't :). But it all comes with cost and complexity.

Even if some of us can play with LSPs easy I would not state that we
should force everyone else into that art.


> Episode 1: - LFA
>   is not deployable due to the algorithm not
>   consider what is an operational feasible
>   path (think of backup of a oc-768 with a oc-3).

If someone has such mix of active core links I would not worry about
LFA, but about redesign of the network first :)


>   so we need:
>
> Episode 2: - LFA + policy (manageability draft)
>
>   trouble is now we have eliminated many paths
>   due to policy reasons: so we need something
>   to increase path-diversity by reusing
>   existing tunnels - which gets us to

As you know both can be automated quite easily.


> Episode 3: - R-LFA
>   it is still possible to construct pathologic
>   topologies which do not provide 100% coverage.

Is the goal to address pathological cases which probability of
happening in real life is equal to any of two of us visiting Mars any
time soon ...

> now my question is: would you bet the farm on that there won't
> be another episode (i.e. things we haven't considered so far).
> i won't, given that track record of misguided hopes and believes.

I am not sure what's the point of that list. If you are stating that
with hop by hop LSP crafting you have more control - of course you are
rigth and no one will argue.

But do you need such fine art or you are ok with lighter and less
expensive alternatives.

Frankly I like control plane based solutions vs data plane .. it is
just that perhaps we should more focus on making the former much
easier for operators then they are today.

Cheers,
R.

_______________________________________________
spring mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/spring

Reply via email to