RE: [802.3ae] Wan Interface Sublayer
Gerry,
I think that this branch of the thread is different than the rest....
I think that the probability that the committee would endorse such a
direction -- especially at this late a stage in the process -- is
extraordinarily unlikely. Even so, there is no reason why a company couldn't
produce such a part and reference both specifications. If it is a good idea
(no cost penalty; broader market potential; etc), it will do well. It
wouldn't be the first time that a company built a part that was a
combination of multiple, standard-based specifications.
jonathan
> -----Original Message-----
> From: Gerry Pepenella [mailto:gjp@xxxxxxxxxx]
> Sent: Friday, August 24, 2001 7:19 AM
> To: Jonathan Thatcher; stds-802-3-hssg@xxxxxxxx
> Cc: Roy Bynum; Ayers, Mike
> Subject: RE: [802.3ae] Wan Interface Sublayer
>
>
> Jonathan,
> Since the interfaces exist for OC192, would it make
> sense to require the
> limits (and test procedure) for 10GBASE-W to be a superset of
> 802.3ae and
> OC192? This would create an interface that could be used with
> both systems
> and I do not believe that there are mutually exclusive
> requirements. The
> change could be simply to reference the GR-1244-CORE and
> GR-499-CORE (and
> corresponding ITU) documents for 10GBASE-W.
> I believe the change would affect only systems that
> wanted to implement
> 10GBASE-W, but this would create a definitive Ethernet over
> SONET standard.
> Regards,
> Gerry Pepenella
>
> -----Original Message-----
> From: Jonathan Thatcher
> [mailto:Jonathan.Thatcher@xxxxxxxxxxxxxxxxxxxx]
> Sent: Thursday, August 23, 2001 6:49 PM
> To: stds-802-3-hssg@xxxxxxxx
> Cc: 'Gerry Pepenella'; Roy Bynum; Ayers, Mike
> Subject: RE: [802.3ae] Wan Interface Sublayer
>
>
> Gerry,
>
> This is a bit too complicated to go into detail here. Let me
> give you the
> 30k-ft version.
>
> The entire jitter specification and test methodology for
> SONET is different
> than in 802.3ae. This is due, in part, to the objectives for the two
> standards organizations and the history of the groups. In the
> case of SONET,
> a primary objective is to limit jitter propagation (yes, this is a
> simplification); in 802.3ae, the primary objective is plug
> and play (no
> engineered links; no jitter propagation).
>
> In fact, while there is certain similarity between the 802.3ae and the
> 802.3z (or Gigabit Fibre Channel) jitter specifications and test
> methodologies, even these are significantly different
> (compare clauses 38
> and 52). Even if the specifications (the actual numbers) were
> the same, it
> wouldn't mean anything since what is measured is not the same.
>
> jonathan
>
> Jonathan Thatcher
> Principal Engineer, World Wide Packets
> Chair, IEEE P802.3ae Task Force
> Office: 509.242.9228 Fax: 509.242.9001
> jonathan@xxxxxxx
>
>
> > -----Original Message-----
> > From: Gerry Pepenella [mailto:gjp@xxxxxxxxxx]
> > Sent: Thursday, August 23, 2001 12:28 PM
> > To: Roy Bynum; Ayers, Mike; stds-802-3-hssg@xxxxxxxx
> > Subject: RE: [802.3ae] Wan Interface Sublayer
> >
> >
> >
> > Roy,
> > The recent clock tolerance changes have aligned OC192
> > with 802.3ae for
> > clock accuracy and I do not believe there is an inherent
> > affect on jitter.
> > There is a difference in the jitter budgets for OC192 and
> > 802.3ae. A quick
> > summary is that the transmit jitter budget for SONET/SDH is
> > tighter than
> > 802.3ae and I believe the receive tolerance is larger for
> > 802.3ae than for
> > SONET/SDH. Since it is desired that the 802.3ae 'W'
> > interfaces are directly
> > compatible with the SONET/SDH network would it make some to
> > align the 'W'
> > interface in all parameters?
> >
> > Regards,
> > Gerry Pepenella
> >
> > -----Original Message-----
> > From: owner-stds-802-3-hssg@xxxxxxxxxxxxxxxxxx
> > [mailto:owner-stds-802-3-hssg@xxxxxxxxxxxxxxxxxx]On Behalf Of
> > Roy Bynum
> > Sent: Thursday, August 23, 2001 8:59 AM
> > To: Gerry Pepenella; Ayers, Mike; stds-802-3-hssg@xxxxxxxx
> > Subject: RE: [802.3ae] Wan Interface Sublayer
> >
> >
> >
> > Gerry,
> >
> > I do not know how the clock tolerance change may have changed
> > the jitter
> > requirements. I do not remember any change to the jitter
> > specifications
> > that were a result of the clock tolerance change. Do you
> > think that there
> > should be?
> >
> > Thank you,
> > Roy Bynum
> >
> > At 08:29 AM 8/23/01 -0500, Gerry Pepenella wrote:
> > >Roy,
> > > Does this apply to jitter requirements as well?
> > >
> > >Gerry Pepenella
> > >Silicon Laboratories
> > >
> > >-----Original Message-----
> > >From: owner-stds-802-3-hssg@xxxxxxxxxxxxxxxxxx
> > >[mailto:owner-stds-802-3-hssg@xxxxxxxxxxxxxxxxxx]On Behalf
> > Of Roy Bynum
> > >Sent: Tuesday, August 21, 2001 11:42 PM
> > >To: Ayers, Mike; 'stds-802-3-hssg@xxxxxxxx'
> > >Subject: RE: [802.3ae] Wan Interface Sublayer
> > >
> > >
> > >
> > >Mike,
> > >
> > >As of Draft 3.1, the clock tolerances of the WAN PHY
> > transmitter are the
> > >same as SONET and SDH Class B SRE and LRE receivers.
> > >
> > >Thank You,
> > >Roy Bynum
> > >
> > >At 04:16 PM 8/21/01 -0500, Ayers, Mike wrote:
> > >
> > >
> > > > > From: Roy Bynum [mailto:rabynum@xxxxxxxxxxxxxx]
> > > > > Sent: Monday, August 20, 2001 07:15 AM
> > > >
> > > > > As for being able to transparently carry 10GbE WAN PHY over
> > > > > an optical
> > > > > network, no additional systems other than what already
> > exists in the
> > > > > service provider infrastructure is required. Only if
> > the WAN PHY is
> > > > > multiplexed into another transmission system other than 10G
> > > > > SONET or SDH is
> > > > > additional infrastructure needed, which is required
> > regardless of the
> > > > > protocol or PHY.
> > > >
> > > > So the clock and tolerances for WIS and SONET are
> > the same.
> > > >Correct?
> > > >
> > > >
> > > >/|/|ike
> >
> >
>