[Nfd-dev] [Operators] monitoring data plane

[Junxiao Shi]

Hi Jeff

http://yoursunny.com/p/ndn6/reachability/ is a multi-HUB test.
Both programs have the same drawback: measured delay includes not only the
delay between HUBs, but also the delay from laptop to connected HUB.

In making this multi-HUB reachability test, I have the following experience:

   - Do not send ping as first Interest, because its measured delay may
   contain the time spent to establish WebSocket connection. Instead, send a
   dummy Interest (eg. "ndn:/"), and send the first ping after the dummy
   Interest is satisfied or timed out.
   - Do not send too many Interests at the same time. Delay is
   significantly increased if rate is more than 5 Interests/second, due to CPU
   bottleneck (on a Windows 7 laptop). Chrome's CPU usage should be kept under
   2% for measured delay to be accurate.
   - ndnpingserver must use RSA signature. SHA256 digest signatures cannot
   be decoded by ndn-js (Feature #1851
   <http://redmine.named-data.net/issues/1851>).


Yours, Junxiao

On Mon, Aug 18, 2014 at 7:10 PM, Burke, Jeff <jburke at remap.ucla.edu> wrote:

>
>  Also, in case it's helpful, I've updated (hastily) the ndn-ping example
> in ndn-js to work with the current ping server and NFD.
> https://github.com/named-data/ndn-js/tree/master/examples/ndnping
>
>  A live version is here:
>  http://named-data.net/apps/live/ndn-ping.html
>
>  This only connects to a single hub, but perhaps it would be useful in
> building a multi-hub test, though the code is not all that well written
> (sorry).
>
>  Jeff
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/nfd-dev/attachments/20140820/437f3c1e/attachment.html>