[ndnSIM] Obtaining CS metrics

sai gautams saigautam96 at gmail.com
Sun Mar 6 18:06:44 PST 2022


Hi Alex,

 I would like to follow your 1st approach where you suggested that I attach
the callback to a specific node's forwarder. Can you please explain to me
briefly how I can do that?


On Sun, Mar 6, 2022 at 9:14 PM sai gautams <saigautam96 at gmail.com> wrote:

> Hi Alex,
>
> Thanks for your explanation. I will try this out.
>
> On Sun, Mar 6, 2022 at 10:12 AM Alex Afanasyev <aa at cs.fiu.edu> wrote:
>
>>
>> On Mar 5, 2022, at 6:02 PM, sai gautams via ndnSIM <
>> ndnsim at lists.cs.ucla.edu> wrote:
>>
>> Hi everyone,
>>
>> I'm trying to obtain the CS metrics. I have installed the ndn::Cs Tracer
>> on all my nodes and set a non-zero CS size for the nodes but the output
>> .txt file is displaying Zeroes. I have found a mailing list where this
>> question was asked previously. The solution provided is CS tracer doesn't
>> support NFD's case which is beyond my understanding. Can anyone please help
>> me in making me understand what exactly should I do? Thanks.
>>
>>
>> Well.. That tracer had limited functionality and was depending on legacy
>> function of very old content store implementation (and wasn't very fast).
>> Cache "metrics" is quite broad term and there isn't really a fit-all
>> solution.
>>
>> It would be great if you can work on a replacement function in ndnSIM, we
>> would be happy to merge your patches.
>>
>> A few ideas.
>>
>> Forwarder class has two signals:  "afterCsHit" and "afterCsMiss".  One
>> way you can get metrics right away, is to simply attach your callback to a
>> specific node's Fowarder and then capture whatever you need.
>>
>> Another way is to export this signals to NS-3 native side into
>> ndn::L3Protocol (model/ndn-l3-protocol.cpp). There are already a few
>> signals converted to NS-3's TraceSources and you can try to follow the
>> idea.   After that, you just need to create/update the tracer with the
>> metrics of your choice and attach to NS-3's trace source, similar how other
>> tracers are attached (or directly from the scenario, also a possibility).
>>
>> -
>> Alex
>>
>>
>>
>
> --
>
> *Sai Gautam Mandapati*
>
> Deakin University
>
>
>

-- 

*Sai Gautam Mandapati*

Deakin University
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/ndnsim/attachments/20220307/7aa10760/attachment.html>


More information about the ndnSIM mailing list