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

RE: Clock Tolerance and WAN PHY


You bring up an interesting point.  Everyone I seem to talk to believes that
in some way, shape or form, the WAN PHY is going to use the SONET/SDH
infrastructure.  A lot of people seem to desire the full SONET overhead and
SONET clock tolerances.  If that is true, then the question comes to my
mind: could we not simplify the WIS?  The WIS would just be a reference to
ANSI T1.416 (or some other SONET standard).  The Task Force would have met
their objectives and would save ourselves a lot of work.

On the flip side, I would think that would be a shame.  I like having a very
lean SONET overhead, loose clock tolerance WAN PHY.  I like the ability to
provide that WAN PHY past the front side piece of equipment.  The more that
the WAN PHY approaches SONET, I personally believe that it is less likely to
see deployment in the network past the bridges (or front side).

Maybe the Task Force should really revisit why they want the WAN PHY.  It
would be a shame to provide something that no one plans on using or that
anyone finds useful.


-----Original Message-----
From: Savara, Raj [mailto:RSavara@xxxxxxxxxxxxxxxxxxx]
Sent: Tuesday, January 23, 2001 9:44 AM
To: 'Tom Alexander'; 'James Colin'; Luigi.Ronchetti@xxxxxxxxxxxxxxxx;
Cc: stds-802-3-hssg@xxxxxxxx
Subject: RE: Clock Tolerance and WAN PHY

Tom and James,

This does not stop someone from specifying a product which will interface
with SONET.  As you pointed out, the optics, clock and OH bytes are all
within the IEEE spec, and can interoperate with SONET equipment.  The only
drawback is interoperating with another IEEE product which does not carry
the OH, which would drop information.  

Personally, I believe the IEEE should adopt the full SONET OH and clock
tolerances.  This makes more sense, and does not cost much more.  The
difference in complexity on the WIS is minimal, and the clock tolerances
were discussed earlier with minimal cost increases.  Most people agreed that
the volume of the 10GE would drive the cost of the crystals down.  The
optics will be in a similar situation.  In fact, the jitter on the IEEE may
be more stringent than ITU G691, or Bellcore.  This means any manufacture
who builds 10GE will have no problems with SONET.

Raj Savara
Network Elements