[Nfd-dev] the impact of cs_size on NFD tool,chunks

Klaus Schneider klaus at cs.arizona.edu
Mon Mar 12 19:22:07 PDT 2018



On 12/03/18 18:19, 小杰许 wrote:
> The topology is like  two clients(without cs)---one NFDrouter(with 
> cs)----server(without cs).

What's the RTT?

> 
> And I set the cs_size of NFDrouter from 1000 to 5000. 
> (1000,2000,3000,4000,5000).

Is that number of packets?

How many chunks is your video file? It might not be much more than 1000.

> 
> the command on the server is
> 
> ndnputchunks -v /ndn < /home/b/panda.avi
> 
> (here panda.avi is a clip of video, it's 20Mb)
> 
> the command on the client is
> 
> ndncatchunks -v /ndn.

Maybe remove the "-v". The console output might be the performance 
bottleneck.

> 
> 
> 
> I run one of the consumer first, after it finished, I run another one.
> 
> the goodput of the first consumer is 5Mbit/s, the second one is random, 
> sometimes higher than 5Mbit/s,and sometimes lower than 5Mbit/s. I think 
> with the cache in the middle NFDtouter, the goodput of the second client 
> could be better all the time, but it's not.
> 
> And with the increase of the cs_size, there is not linear change of the 
> clients' goodput.
> 
> BTW, the ndncatchunks in the collapse more than half time when I run, 
> and with messy code (cause the file is a video?) in the terminal. With 
> the file (.txt), I got ERROR (retry(3).........).

Please copy the summary output here. Also for the runs that worked.

Best regards,
Klaus

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


More information about the Nfd-dev mailing list