[Nfd-dev] RTT delay caused by RIB registration commands

Vince Lehman (vslehman) vslehman at memphis.edu
Wed May 6 16:04:45 PDT 2015


At the NFD call today, it was suggested that I try our experiments with a different, quicker signing method for NFD’s management modules.

I ran the following three tests:

1.) Unmodified NFD to get baseline results
2.) NFD with management modules using KeyChain::signWithSha256 for control responses
3.) Test 2 + modified ndn::nfd::Controller in ndn-cxx so control commands are signed using KeyChain::signWithSha256

The results I collected suggest that signing is largely the cause of the RTT jumps.

1. ) Unchanged NFD using KeyChain::sign():

20150506T220013.546312 - Content From /ndn/edu/orange - Ping Reference = 358723384   - Round Trip Time = 7.54174 ms
20150506T220014.545910 - Content From /ndn/edu/orange - Ping Reference = 358723385   - Round Trip Time = 7.35931 ms
20150506T220015.549327 - Content From /ndn/edu/orange - Ping Reference = 358723386   - Round Trip Time = 10.755 ms
20150506T220016.569095 - Content From /ndn/edu/orange - Ping Reference = 358723387   - Round Trip Time = 26.909 ms
20150506T220017.554450 - Content From /ndn/edu/orange - Ping Reference = 358723388   - Round Trip Time = 15.8414 ms
20150506T220018.550272 - Content From /ndn/edu/orange - Ping Reference = 358723389   - Round Trip Time = 11.7576 ms
20150506T220019.594078 - Content From /ndn/edu/orange - Ping Reference = 358723390   - Round Trip Time = 52.0413 ms
20150506T220020.572899 - Content From /ndn/edu/orange - Ping Reference = 358723391   - Round Trip Time = 31.8473 ms
20150506T220021.568606 - Content From /ndn/edu/orange - Ping Reference = 358723392   - Round Trip Time = 29.9206 ms
20150506T220022.545620 - Content From /ndn/edu/orange - Ping Reference = 358723393   - Round Trip Time = 7.09895 ms
20150506T220023.551988 - Content From /ndn/edu/orange - Ping Reference = 358723394   - Round Trip Time = 13.4654 ms
20150506T220024.564711 - Content From /ndn/edu/orange - Ping Reference = 358723395   - Round Trip Time = 26.1983 ms
20150506T220025.590458 - Content From /ndn/edu/orange - Ping Reference = 358723396   - Round Trip Time = 51.8421 ms
20150506T220026.547079 - Content From /ndn/edu/orange - Ping Reference = 358723397   - Round Trip Time = 8.38067 ms
20150506T220027.546382 - Content From /ndn/edu/orange - Ping Reference = 358723398   - Round Trip Time = 7.7887 ms

2.) NFD Management using KeyChain::signWithSha256() for control responses:

20150506T220832.990110 - Content From /ndn/edu/orange - Ping Reference = 481750786   - Round Trip Time = 8.12952 ms
20150506T220833.989616 - Content From /ndn/edu/orange - Ping Reference = 481750787   - Round Trip Time = 7.27718 ms
20150506T220835.000043 - Content From /ndn/edu/orange - Ping Reference = 481750788   - Round Trip Time = 17.7199 ms
20150506T220835.997854 - Content From /ndn/edu/orange - Ping Reference = 481750789   - Round Trip Time = 15.8593 ms
20150506T220836.990831 - Content From /ndn/edu/orange - Ping Reference = 481750790   - Round Trip Time = 8.84247 ms
20150506T220838.011070 - Content From /ndn/edu/orange - Ping Reference = 481750791   - Round Trip Time = 17.0122 ms
20150506T220839.022594 - Content From /ndn/edu/orange - Ping Reference = 481750792   - Round Trip Time = 40.5401 ms
20150506T220839.996497 - Content From /ndn/edu/orange - Ping Reference = 481750793   - Round Trip Time = 13.9247 ms
20150506T220841.000317 - Content From /ndn/edu/orange - Ping Reference = 481750794   - Round Trip Time = 18.0118 ms
20150506T220841.990391 - Content From /ndn/edu/orange - Ping Reference = 481750795   - Round Trip Time = 8.36782 ms
20150506T220842.989631 - Content From /ndn/edu/orange - Ping Reference = 481750796   - Round Trip Time = 7.57163 ms
20150506T220843.990083 - Content From /ndn/edu/orange - Ping Reference = 481750797   - Round Trip Time = 7.9309 ms

3.) NFD Management using KeyChain::signWithSha256() and ndn::nfd::Controller::start() using KeyChain::signWithSha256:

20150506T224746.920419 - Content From /ndn/edu/orange - Ping Reference = 432571709   - Round Trip Time = 7.74707 ms
20150506T224747.920684 - Content From /ndn/edu/orange - Ping Reference = 432571710   - Round Trip Time = 8.08427 ms
20150506T224748.921073 - Content From /ndn/edu/orange - Ping Reference = 432571711   - Round Trip Time = 7.95967 ms
20150506T224749.920619 - Content From /ndn/edu/orange - Ping Reference = 432571712   - Round Trip Time = 7.78507 ms
20150506T224750.923189 - Content From /ndn/edu/orange - Ping Reference = 432571713   - Round Trip Time = 10.5592 ms
20150506T224751.920660 - Content From /ndn/edu/orange - Ping Reference = 432571714   - Round Trip Time = 7.73255 ms
20150506T224752.920939 - Content From /ndn/edu/orange - Ping Reference = 432571715   - Round Trip Time = 8.24211 ms
20150506T224753.923819 - Content From /ndn/edu/orange - Ping Reference = 432571716   - Round Trip Time = 9.19251 ms
20150506T224754.921798 - Content From /ndn/edu/orange - Ping Reference = 432571717   - Round Trip Time = 8.94248 ms
20150506T224755.924729 - Content From /ndn/edu/orange - Ping Reference = 432571718   - Round Trip Time = 12.0222 ms
20150506T224756.920302 - Content From /ndn/edu/orange - Ping Reference = 432571719   - Round Trip Time = 7.75629 ms


--
Vince Lehman

On May 5, 2015, at 8:37 AM, Vince Lehman (vslehman) <vslehman at memphis.edu<mailto:vslehman at memphis.edu>> wrote:

Hi Junxiao,

A registration for a new UDP face causes the RTT to increase slightly, ~2ms above the previous experiment:

Content From /ndn - Ping Reference = 477477776   - Round Trip Time = 0.778 ms
Content From /ndn - Ping Reference = 477477777   - Round Trip Time = 0.649 ms
Content From /ndn - Ping Reference = 477477778   - Round Trip Time = 16.614 ms
Content From /ndn - Ping Reference = 477477779   - Round Trip Time = 0.982 ms

A registration for a non-existent face ID causes a large increase in the RTT:

Content From /ndn - Ping Reference = 479746716   - Round Trip Time = 0.77 ms
Content From /ndn - Ping Reference = 479746717   - Round Trip Time = 0.695 ms
Content From /ndn - Ping Reference = 479746718   - Round Trip Time = 52.996 ms
Content From /ndn - Ping Reference = 479746719   - Round Trip Time = 0.808 ms

Since we simply want to update the expiration time for a particular namespace in the RIB, would there be any use for a protocol extension that allows for the expiration time for a namespace to be updated with one RIB command?
--
Vince Lehman

On Apr 24, 2015, at 1:38 PM, Junxiao Shi <shijunxiao at email.arizona.edu<mailto:shijunxiao at email.arizona.edu>> wrote:

Hi Lan

Before I can suggest any code patches, additional experiment is needed to confirm or reject the hypothesis in http://www.lists.cs.ucla.edu/pipermail/nfd-dev/2015-April/001049.html .

If you want to measure RTT changes due to route changes, one possible way is to increase the link delay.
When the link delay is 10ms, a 10ms RTT change is 50% and appears significant.
When the link delay is 500ms, a 10ms RTT change is 1% and appears negligible.

Yours, Junxiao

On Fri, Apr 24, 2015 at 11:33 AM, Lan Wang (lanwang) <lanwang at memphis.edu<mailto:lanwang at memphis.edu>> wrote:
Basically, we saw a large jump in ndnping results between two nodes even though there was no path change for the path between those two nodes.  We found that it was because those nodes (or the intermediate ones) were busy updating their FIB (for other routes) -- each registration and unregistration of a route obviously took ~10ms for NRD/NFD to process.

We were interested in RTT changes due to route changes, so the extra delay caused by the route registrations is an undesirable side effect we'd like to remove (thus Vince's email).  So any suggestions on how to deal with this in our experiments are highly appreciated.

In the long term, we do need to address the problem of high processing delays of route registrations.

Lan


_______________________________________________
Nfd-dev mailing list
Nfd-dev at lists.cs.ucla.edu<mailto:Nfd-dev at lists.cs.ucla.edu>
http://www.lists.cs.ucla.edu/mailman/listinfo/nfd-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/nfd-dev/attachments/20150506/ee313c27/attachment.html>


More information about the Nfd-dev mailing list