[Nfd-dev] What is the desired behavior of Faces to dead nodes?

Syed Obaid Amin obaidasyed at gmail.com
Mon Nov 24 09:21:19 PST 2014


Hi Junxiao,

Thanks for the clarification.

But may I know the rationale behind this design? Like can't we ignore the
ICMP error for UDP tunnels? Closing a face makes sense for stateful
connections, but I think for stateless connections like UDP, if the remote
end goes down the local end should not be affected?

Regards,
Obaid

On Mon, Nov 24, 2014 at 10:59 AM, Junxiao Shi <shijunxiao at email.arizona.edu>
wrote:

> Hi Obaid
>
> This behavior is by design.
>
> The face created is a persistent face, which would not expire on idle, but
> would be closed if it fails.
> When something is sent out of the face to a non-existent port, ICMP error
> comes back, which causes the face to fail and be closed.
> When a face is closed, Routes and FibNextHop records associated with that
> face are removed.
>
> Yours, Junxiao
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/nfd-dev/attachments/20141124/e262387f/attachment.html>


More information about the Nfd-dev mailing list