[ndnSIM] Memory leaks

Philipp Moll philipp.moll at itec.aau.at
Tue Jan 31 23:54:05 PST 2017


I also recognized this memory leak. In some cases the memory consumption 
of my simulations increase linear over the time; at the end of the 
simulation they consume up to 60GB. This problem especially arises in 
scenarios where Multicast is used as FW-Strategy and many packets are sent.

I’m currently working with ndnSIM-2.1 (Commit-Hash: a9d889b).

Philipp Moll

On 31/01/17 19:14, Saad SA wrote:
> No stabilization, it keeps growing up till the end of the simulation 
> (Yes, it finishes). So if the same simulation ran on two different 
> computers; the faster one would finish before, hence consuming less 
> memory than the slower, the growing stops just when the simulation 
> ends. (You have a rate of growing equal to X MB/s, so taking long time 
> to finish would lead to consume more!)
> Sa'd
> On Tue, Jan 31, 2017 at 7:56 PM, Alex Afanasyev <aa at cs.ucla.edu 
> <mailto:aa at cs.ucla.edu>> wrote:
>     > On Jan 31, 2017, at 8:31 AM, Saad SA <maxpromindz at gmail.com
>     <mailto:maxpromindz at gmail.com>> wrote:
>     >
>     > Hi,
>     >
>     > I had an application on ndnSIM version 0.2.8, and it was working
>     too fine. Now I'm using the latest version, I changed the old
>     application to work over the new ndnSIM and noticed huge memory
>     leaks that are proportional to the run-time; On some computer it
>     took more time and consumed 10 GB of RAM !! and on another faster
>     one, it took less, hence consumed less RAM (4 GB or less), both
>     are still huge amounts of memory as I'm using a small topology
>     with 18 nodes and an average frequency of 500 requests/s; and all
>     of that was running fine on the older version of ndnSIM.
>     > Any ideas?
>     Can you clarify a bit (if you can share a simple scenario that
>     highlights the issue, it would be even better). Was the memory
>     consumption growing or got to a point of high utilization and
>     stabilized (you said it is run-time, but then mentioned that
>     simulation finished).  The reason I'm asking is that the old
>     version of ndnSIM had optimization of "virtual payloads", which is
>     not available in the new version.  This has impact on the amounts
>     of memory used, but it should be just a constant factor increase,
>     not continuous increase during runtime.
>     --
>     Alex
> _______________________________________________
> ndnSIM mailing list
> ndnSIM at lists.cs.ucla.edu
> http://www.lists.cs.ucla.edu/mailman/listinfo/ndnsim
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/ndnsim/attachments/20170201/79a93bda/attachment-0001.html>

More information about the ndnSIM mailing list