Segment Routing is as much source routing as RSVP-TE would be, there’s no per
hop SA lookup, but rather per FEC (that could be a SA) at the ingress.
Ability to abstract next hop into higher layer, locally resolvable entity
(adj-sid -> node->sid -> anycast-sid -> etc) provides a rather viable from a
network design and implementation way to have end2end networking while
utilizing local state.
Cheers,
Jeff
On 12/31/16, 04:22, "ietf on behalf of Stewart Bryant"
<ietf-bounces(_at_)ietf(_dot_)org on behalf of
stewart(_dot_)bryant(_at_)gmail(_dot_)com> wrote:
On 31/12/2016 07:54, Octavio Alvarez wrote:
> On 12/30/2016 02:20 PM, Masataka Ohta wrote:
>
>> I wrote:
>>
>> https://tools.ietf.org/html/draft-ohta-e2e-multihoming-00
>>
>> in April 2000 and I know it is stupid to use source routing for
>> multihoming.
> Is source routing bad in general or is it bad only in multihoming
scenarios?
>
> Thanks.
>
When called segment routing it is accepted as a useful technology for
well defined problems of this type.
- Stewart