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

RE: Clause 33 comments






Boaz,
Thank you for reviewing the Clause. Your first point ties into one of my big
issues  :
Where is link status reported ?

What I was alluding to in point 12 was a possible implementation where there is
just the DTE XGXS, a re-timer (on both tx and rx) and the WWDM PMD.  Such an
implementation was proposed on slide 9 of hanson_1_1199 in the November 99
presentation archive
(http://www.ieee802.org/3/10G_study/public/nov99/hanson_1_1199.pdf).

There are two options for the 're-timer' block:
1/ logically decode the serial streams to parallel data and then re-encode back
to serial.  This is the same as a PHY XGXS device sat on top of a 10GBASE-X PCS
layer. In this system the 10GBASE-X PCS exists as a chunk of logic and it can
provide information via MDIO on all the things like link status.
2/ do just a clock recovery and re-timing function with no logical decode.  In
this case, the four incoming data streams  are not decoded until they reach the
DTE XGXS.  There is no 10GBASE-X PCS and so the DTE XGXS is effectively the PCS.

The best I've been able to come up with so far for link status goes as follows :
All implementations except the re-timed WWDM have a PCS device which actively
determines link status and can report it back via MDIO.  However, in systems
with the optional XAUI interface, the status of the PCS link OK needs to be
qualified with the link status of XAUI.  For such systems, it would be possible
to have the DTE XGXS report the overall link status by getting the PCS to signal
remote fault up to the DTE XGXS whenever its own link status dropped.  The STA
management would be able to interrogate the PCS to check if it was generating
the remote fault condition or forwarding it on from the far end of the link.
For the re-timed WWDM case, the re-timer could respond as a 10GBASE-X PCS
device, indicating to management that it is a re-timer and link status needs to
be checked in the DTE XGXS.

To answer your second point, when D1.0 was issued, 1310nm WWDM and 850nm serial
had not been approved by the Task Force and so I omitted them from the
registers.  I shall be updating the tables for the next revision to cover these
new additions.

Regards,
Ed





Boaz Shahar <boazs@xxxxxxxxxxxx> on 04/10/2000 16:50:32

Sent by:  Boaz Shahar <boazs@xxxxxxxxxxxx>


To:   Edward Turner/GB/3Com@3Com, stds-802-3-hssg@xxxxxxxx
cc:
Subject:  RE: Clause 33 comments






> -----Original Message-----
> From: Edward Turner [mailto:Edward_Turner@xxxxxxxxxxxx]
> Sent: Tuesday, October 03, 2000 4:32 PM
> To: stds-802-3-hssg@xxxxxxxx
> Subject: Re: Clause 33 comments
>

>
> 12/ This is a big issue.  The PCS registers at the moment
> just cover the 64/66
> codec.  The question is whether the WWDM PCS is covered by
> the DTE XGXS device ?
>

I want to comment about this one-The WWDM PCS is not covered by the DTE XGXS
since the later does not aware to whether it is connected to WDM/Serial LAN
PHY or Serial WAN PHY (The XAUI may I/F any of them)

And in the same issue, in table 33-5, the 4x3.125 parallel and WDM options
are not mentioned.

Regrads,
Boaz



>