I guess there are some potential advantages in deploying multipath
inter-domain routing. For instance, currently each border router
forwards the traffic for a prefix through the same next AS. Balancing
through two different next ASes it is not possible without re-
advertising and sending the control plane to reconverge. Using
multipath the balancing could be done even in a fine granularity (e.g.
per packet) so that,
- Different treatment (e.g. regarding QoS) could be provided per
packet/flow, relaying traffic to a better provider (likely to have a
higher pricing) and charging more money to the customers.
- The latter may enable pricing and services competition.
- Egress traffic engineering could be decoupled from the routing
advertisement, forwarding to an upstream provider or another should be
transparent to downstream customer ASes if they all the paths are
advertised beforehand.
- Congestion control may be more flexible for ISPs.
Thinking of future architectures, the case of multipath-aware
transport protocols would let some or full control of the path
selection to final hosts. It is not likely that ISPs go for this, but
theoretically failure detection, failure recovery and better
congestion control should be possible in this scenario, since end-
hosts can measure end-to-end performance easily.
Jose M. Camacho, [email protected], www.uc3m.es
+34916248795
El 11/05/2011, a las 0:36, Sampo Syreeni escribió:
On 2011-05-10, marcelo bagnulo braun wrote:
What about doing some work in multi-path routing in the inter-
domain? i.e. enabling a router to use multiple paths with potential
different characteristics to forward packets across domains. Would
people find this topic interesting?
I'd guess they would, since the idea is so old and also
underexploited. But does it bring in real extra value? I.e. does
anybody know whether multipath routing could significantly increase
the total utilization of the routing fabric, either now or under
some theoretically founded future situation?
If nobody knows, a preliminary study could be IRTF material.
--
Sampo Syreeni, aka decoy - [email protected], http://decoy.iki.fi/front
+358-50-5756111, 025E D175 ABE5 027C 9494 EEB0 E090 8BA9 0509 85C2
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg
