Re: Leveraging OC-192c - 802.3x flow control
Part of your last sentence sums up the real objective at the root of the 10.0
vs. 9.584640 discussion "...we need to keep in mind
the useful throughput we want on the link." I agree with you that the
coexistence of 10.0 Gbps Ethernet and 9.584640 SONET is largely an
implementation choice based on existing architectures.
Hon Wah Chin wrote:
> Rich -
> I was just following up on the discussion of using
> 802.3x for pacing with the observation that
> doing so might interact with allowing 802.3x across
> the link.
> There was a previous comment on that,
> and I was pointing out that using 802.3x at this MAC/PLS
> might mean either disallowing pacing across the link,
> or having extra state within the PLS/PHY/PMD complex.
> My first paragraph suggested that with the MAC/PLS interface
> being in most cases hidden from the outside, 10Gb/s to OC-192
> pacing mechanism need not be 802.3x or even defined in committee,
> so the flow control discussion would be moot.
> (The implementor simply clocks information out of the system
> more slowly. 802.3x is used for the overall link, as before.)
> The more important discussion, to my mind, is to make sure of good
> interoperability at the PHY port WITHOUT too many different versions.
> And in defining the signalling on the link, we need to keep in mind
> the useful throughput we want on the link.
> -----Original Message-----
> From: Rich Taborek [mailto:rtaborek@xxxxxxxxxxxxxxxx]
> Sent: Monday, July 26, 1999 11:23 AM
> To: Hon Wah Chin; HSSG
> Subject: Re: Leveraging OC-192c - 802.3x flow control
> Hon Wah,
> I agree in general with your thoughts on leveraging elements of the OC-192
> infrastructure including EDFAs and regenerators for use with 10 GbE
> The only question I have is with respect to 802.3x flow control: My
> understanding is that 802.3x flow control is limited to a single link. Is
> this your understanding also? I believe we need to clear up this matter as
> continue to try and determine the correct HSSG speed objective.
> A second point you make is to raise the possibility of using 802.3x AND
> other pacing mechanisms in a link. I don't believe that it's a good idea to
> mechanisms that control the flow of packets within a link. This speaks in
> favor of 802.x as being the only flow control mechanism for 10 GbE.
> Best Regards,
Richard Taborek Sr. Tel: 650 210 8800 x101 or 408 370 9233
Principal Architect Fax: 650 940 1898 or 408 374 3645
Transcendata, Inc. Email: rtaborek@xxxxxxxxxxxxxxxx
1029 Corporation Way http://www.transcendata.com
Palo Alto, CA 94303-4305 Alt email: rtaborek@xxxxxxxxxxxxx