[Nfd-dev] best-route behavior

[Junxiao Shi]

Hi Hila

The consumer app needs to retransmit a similar Interest (same
Name+Selectors, different Nonce). The retransmission will be forwarded to
another nexthop (toward D).

best-route v3 has no memory beyond single PIT entry.
This means, even if a retx finds a different producer, a different Interest
still needs consumer retx.
The assumption is: routing protocol is expected to erase the incorrect
route (toward C) shortly.

Yours, Junxiao

On Mon, Apr 20, 2015 at 8:41 AM, Hila Ben Abraham <hila at wustl.edu> wrote:

> All,
>
> What is the expected behavior of the best-route strategy when the primary
> face it uses doesn't respond with data? Shouldn't it try another face?
> According to our experiments it keeps on forwarding the same interests to
> the primary face, although another 'working' and more expensive face is
> available.
>
> In some more details:
> We are running an experiment using the following topology:
> A--B--C
>       |
>      D
>
> On B, we set the cost of /ndn/test/ping toward C to be cheaper than the
> one toward D.
>
> We start ndnpingserver on node D to respond to /ndn/test/ping, and we
> start ndnping to ping for the same namespace on node A.
>
> As expected, B forwards the first interest on the cheapest face toward C.
> However, even though C fails to respond with data, B keeps forwarding the
> following ping interests to C, and never tries the face toward D.
>
> Is this the expected behavior? Am I missing something here?
>
> Thanks,
> Hila
>
> _______________________________________________
> Nfd-dev mailing list
> Nfd-dev at lists.cs.ucla.edu
> http://www.lists.cs.ucla.edu/mailman/listinfo/nfd-dev
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/nfd-dev/attachments/20150420/216f7b36/attachment.html>