Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: [EFM] [EFM-OAM] OAM Transport Proposal





Brad,

I began compiling such a list when I first signed on to
support Hiroshi's OAMiP proposal. I believe there is a
need for this type of OAM for just the types of devices
that Sergiu has mentioned, as well as for low latency
link level messages.

However, the list became very long and involved. I think
the approach we took for 10G, by having the RS provide
the alignment to the PCS by "sliding" the frame +/- 1
byte and using something akin to a Deficit Idle Count,
is probably the easiest approach to make this work. This
results in changes to the transmit RS, PCS and GMII.

Another approach would be to have the PCS pass its
alignment up to the RS somehow. This also results in
changes to each of these sublayers/interfaces but the
cycle to cycle timing would be a little more difficult.

Either way, the changes to the RS aren't too bad because
the odd alignment that results in the discard of the
first byte of the preamble would never happen from
the RS. The capability can remain in the PCS but is
just not used by the RS (when operating in this mode).
The PCS would just have to either provide its even/odd
alignment to the RS or take its alignment from the RS.

We already know how to modify the RS as it can be based
on Clause 46. It will either create its own alignment
or take its alignment from the PCS.

The GMII will be modified to add an additional signal,
preferably in the direction from the RS to the PCS.

I think this is a fair summary but I'm sure there will
be much discussion around it. :)

Regards,
Ben

"Booth, Bradley" wrote:
> 
> Sergiu,
> 
> I assume that an 802.3 repeater would be similar in concept to this
> converter you mention.  For that, I can understand that a MAC level OAM may
> not be sufficient to provide the link/PHY level OAM that would be required.
> 
> As much as OAM in preamble may solve the problem, I am still concerned about
> the changes required to 802.3 to provide this capability.  I think that it
> would be great if an OAM in preamble supporter could compile a list of the
> changes required to existing clauses to add this feature.
> 
> I just want the task force and working group to be well-informed about what
> we're signing up to do.
> 
> Thanks,
> Brad
> 
> ______________________
> Sent from my Blackberry...


-- 
-----------------------------------------
Benjamin Brown
AMCC
2 Commerce Park West
Suite 104 
Bedford NH 03110
603-641-9837 - Work
603-491-0296 - Cell
603-626-7455 - Fax
603-798-4115 - Home Office
bbrown@xxxxxxxx
-----------------------------------------