Re: new draft for data center benchmarking: draft-dcbench-def

Hi Ramki,

Thank you for taking time to provide feedback, much appreciated!

Please see inline

On 6/18/13 7:37 PM, ramki Krishnan wrote:
> Thanks Lucien -- this draft is certainly useful. Please find comments below.
>
>>> For bit forwarding devices: The time interval starting when the end
>     of the first bit of the input frame reaches the input port and ending
>     when the start of the first bit of the output frame is seen on the
>     output port.
>
> Please add the cut-through switching aspect for additional clarity.

This was defined this in section 2.1:

    Latency can be measured with the following methods, irrespectively of
    the type of switching device (bit forwarding aka cut-through or store
    forward type of device)

>
>>> section 7.2., 7.3
>
> Given the wide variation in servers, adapters, OS, Virtualized/non-virtualized  environments etc. it may be hard to benchmark goodput. Perhaps a useful metric to benchmark would be the tcp performance in a pure test environment.

I believe it's better to not get into this specific as we could use either :
-test traffic generator to measure pure test environment
-real application environment to generate traffic

i think it would be better to stay more generic in the beginning and not 
restrict ourselves to pure test environments only
>
>>> 6.1 buffer size
>
> Need to cover buffering per priority (typically 8 priorities) besides port and global.

we have covered the cos and also the dscp values and they should be done 
for port and global in my opinion, hence it was defined globally in 
section 6.1.1



>
> Thanks,
> Ramki
>
> -----Original Message-----
> From: [email protected] [mailto:[email protected]] On Behalf Of Lucien Avramov (lavramov)
> Sent: Wednesday, June 05, 2013 11:07 AM
> To: [email protected]
> Subject: [bmwg] new draft for data center benchmarking: draft-dcbench-def
>
> Hi BMWG,
>
> At IETF 86, Jacob presented the idea for a data center benchmarking definition and methodology. Our goal is to create data center specific benchmarking guidance as this differs in multiple areas from usual general purpose benchmarking.
>
> Thanks to the feedback, we adopted the recommendation to break our work into multiple drafts.
>
> The first draft explaining the concepts and definitions relative to data center benchmarking of network equipment is posted here:
> http://www.ietf.org/id/draft-dcbench-def-00.txt
>
> Please provide non-editorial comments to the list and editorial comments directly to me.
>
> We are looking forward to hear from you.
>
> Cheers,
> Lucien
>
>
> _______________________________________________
> bmwg mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/bmwg
>