[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: [bmwg] WGLC on IPv6 Benchmarking
This would be the third option mentioned in my earlier e-mail: list the
traditional set and then make a note about the smaller possible size and
the reasons for using it. In fact we could even avoid mentioning the
value but rather mention the way to construct it (use just the IPv6
header). I also like this approach.
Best Regards,
Chip
-----Original Message-----
From: Curtis Villamizar [mailto:curtis at occnc.com]
Sent: Tuesday, July 17, 2007 11:12 AM
To: Scott O. Bradner
Cc: bmwg at ietf.org
Subject: Re: [bmwg] WGLC on IPv6 Benchmarking
In message <20070717015253.8F41D42A2BE at newdev.harvard.edu>
Scott O. Bradner writes:
>
>
> David sed:
> > One possible solution: For both Ethernet and Sonet, define two short
> > lengths -- one that's just the minimum L2 framing+IPv6 header to
> > determine throughput and another that includes a signature field to
> > measure latency and sequencing.
>
> I'm not sure Ethernet is a problem unless one has a long signature due
> to Ethernet's mininum packet size but the concept makes sense to me
>
> Scott
Would it be possible/practical to list a set of very small packet sizes
to test to so as to avoid repeating the recurring argument on the WG
mailing about which tiny packet size to test against?
Let the marketeers argue over which one is most meaningful.
Curtis
_______________________________________________
bmwg mailing list
bmwg at ietf.org
https://www1.ietf.org/mailman/listinfo/bmwg
_______________________________________________
bmwg mailing list
bmwg at ietf.org
https://www1.ietf.org/mailman/listinfo/bmwg