[Nfd-dev] two question about NFD

[Lan Wang (lanwang)]

On Jun 27, 2018, at 12:37 PM, Junxiao Shi <shijunxiao at email.arizona.edu<mailto:shijunxiao at email.arizona.edu>> wrote:

Hi Lan

Hi Junxiao and Alex,
Please send any and all technical questions to nfd-dev mailing list, not to individual developers, so that future users can search for answers.

1. suppose an application sends an interest to NFD with the next hop set to NULL.  Before the interest expires, the application sends another interest with the same name but not setting nexsthop to NULL to NFD, will the second interest be sent out or just aggregated with the first interest?
This is an undocumented behavior. It likely depends on the strategy. Consumer should be consistent on whether to use NextHopFaceId, and not rely on NFD behaving one way or another.

2. suppose an application on a node listens on a name prefix but also sends interest out with a name under that name prefix.  If the node receives an interest with the same name, will the interest be delivered to the application?
No. NFD treats the face connected to the application as point-to-point. Forwarding the Interest back to the application would result in a loop, so NFD does not do that.

They are not the same interest though (not the same nonce).  Why is this a loop?

In sync protocols, the same sync interest is pending on both directions on each link. Isn’t this also a loop in your definition?  How does this happen?

Lan

If part of an application acts as a producer and another part of the same application acts as a consumer for that producer's content, ndn-cxx's face has an "Interest loopback" feature (see #3979) that internally invokes the producer callback. This does not go through NFD. Interest loopback feature is turned out by default, but the producer code may turn it off with interestFilter.allowLoopback(false) function call (example in ChronoSync<https://github.com/named-data/ChronoSync/commit/4e010bced3475517bcd2bc48d08743a5afe4f82d>).

Yours, Junxiao

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/nfd-dev/attachments/20180627/7f20cc8e/attachment-0001.html>