ptpd version with kernel send time stamps - synchronization improvements

[gertjan hofman]

We didn't change any of the time stamping mechanism but found that we could
get improvement by taking a closer look at the filtering applied to the
Master to Slave Delay etc.  Especially with store and forward switches,  it
is not obvious to me that the kernel time stamping is the dominating factor
in the time synchronization.

I posted something on this at the old (sourceforge) mailing list... lets
see...  here it is:
http://sourceforge.net/projects/ptpd/forums/forum/469207/topic/1937491

We generally get < 50 uS using ARM embedded boards and a i386 master,  2-4
switches in the loop and > 30Mb/s traffic.  For us the trick was VLAN
prioritization and additional filtering on the master to slave delay, by
rejecting values that were clearly buffered too long by the switches.

Cheers

Gertjan



-- 
==================================================
Gertjan Hofman
ghofman [at] gmail.com   gertjan.hofman [at] honeywell.com

604-982-3574
==================================================
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.infradead.org/pipermail/ptpd/attachments/20100129/85c3a2ff/attachment.htm>