[ndnSIM] Using ndn::Face in ndnSIM environment.

Junxiao Shi shijunxiao at email.arizona.edu
Thu Oct 20 18:16:11 PDT 2016


Hi Anil

When you have two instances of the same application running on two separate nodes, of course you have to use two separate ndn-cxx Face instances, each connected to the NFD instance on the same node.
It’s wrong to share one ndn-cxx Face instance across multiple nodes. This would be similar to sharing the same Unix socket on multiple nodes, and implies there’s an out-of-band communication channel between an application on node A and the NFD on node B. Note: this statement assumes every node is an independent networked device, and does not apply to supercomputers.

My earlier comment <http://www.lists.cs.ucla.edu/pipermail/ndnsim/2016-October/002972.html> addresses the choice between using one face and using two faces on the same node.

Yours, Junxiao

> On Oct 20, 2016, at 6:09 PM, Anil Jangam <anilj.mailing at gmail.com> wrote:
> 
> I agree and to clear the ambiguity, lets say, the "application" is the (class) which derives from "class MyApp : public Application" or from "class MyApp : public App". Now there are two nodes (node-A) and node (node-B) with 'MyApp' installed on it. So if I understood you correctly, its better using different faces from each 'module' interacting with other 'module' within same application or with another application altogether (its rather more important when they are using overlapping name prefixes). I guess we do not capture this aspect in the porting guide and perhaps worthwhile to mention it. Can you please check? 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/ndnsim/attachments/20161020/fa43d366/attachment.html>


More information about the ndnSIM mailing list