[Nfd-dev] NLSR/NFD Issue on the Testbed

[Dehart, John]

Junxiao,

Thanks for your help on this.
I went ahead and entered a redmine issue for it and added it to the list to be discussed tomorrow on the platform call.

https://redmine.named-data.net/issues/4874

John


On Mar 12, 2019, at 11:06 AM, Junxiao Shi <shijunxiao at email.arizona.edu<mailto:shijunxiao at email.arizona.edu>> wrote:

Hi John


For the first Interest which comes from NLSR, if I understand right, NLSR includes information to force this
interest out the face that the LSA came in on. There is no FIB entry yet for /ndn/edu/memphis/…

Now this starts to make sense. There's no FIB entry during both Interests. The first Interest was forwarded with NextHopFaceId.

When the Interest then comes in from BASEL, there is still no FIB entry for the name /ndn/edu/memphis/…  , right?

The second Interest was Nacked by ASF due to no FIB entry, i.e. the first condition.

So, is the lack of a FIB entry at this point causing the NACK? And if so, why could that NACK also clear the PIT entry for the NLSR sent Interest?

Nack normally does not delete the PIT entry, since it's still waiting on face 278 to respond. ASF explicitly deletes the PIT entry by invoking Strategy::rejectPendingInterest, because it thinks the Interest cannot be satisfied.
This indicates ASF cannot effectively accommodate a runtime strategy change: it would delete a PIT entry after another strategy (or NextHopFaceId in this case) has forwarded the Interest.
I'll leave this to ASF designers to decide whether this is a strategy bug.

Yours, Junxiao

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/nfd-dev/attachments/20190312/e3832cd8/attachment.html>