Re: [EFM][EFM-P2P] 100Mbps P2P over SMF
Hallo Ulf,
I am also interested to discuss 100M specs for an active architecture.
I am very sorry that I missed your talk in the last Austin meeting (I cancelled
my trip after the crash in Queens, NYC on Monday Nov. 12). Is the presentation
on the web the one that you presented in Austin?
Many greetings from Dora
"Horne, David M" wrote:
> Hi Ulf, I would be interested in 100M, particularly in the context of the
> local legs of active architectures and the HALF-PON architecture I brought
> up a while back (http://www.ieee802.org/3/efm/public/email/msg00357.html ).
> It has some similarities to your recent presentation at:
> http://grouper.ieee.org/groups/802/3/efm/public/nov01/jonsson_1_1101.pdf
>
> We will run into the same multiple-distance-support questions as 1G P2P.
> 100M seems most interesting to me--in the context of an active
> architecture--with a long distribution fiber at 1G, and short feeders on the
> local legs to each ONU at 100M (or, 1G broadcast-and-select passive
> downstream, and 100M upstream to an active node for HALF-PON). How short is
> short, is the question. Maybe 300m; maybe 500m; maybe 1-2km. Inexpensive is
> key. The INTENT should be to serve short feeders in this range, rather than
> trying to be all things to all people. The distance extension premium may be
> too high to justify the cost per user of the longer distance capability as
> an umbrella solution.
>
> If the network layout needs longer multi-km feeders, a different
> architecture (or proprietary distance-extended 100M) would be used. 300m
> serves a high percentage of medium- to high-density residential applications
> at 16:1 aggregation. Possibly (haven't run any numbers on this) 64:1 fits
> with 1km. If the cost to increase from 300m to 1km is only a few bucks per
> user, then it makes sense for 1km to be the goal. If it costs, say, 30 bucks
> to go from 300m to 1-2km capability, it will be hard to justify anything but
> 2 separate specs. 30 bucks times zillions of users is a lot of Cap-ex that
> goes wasted for a network operator whose layout logistics dictate deployment
> at 16:1 / 300m, but has to pay the premium for the capabilities to reach 1km
> because the spec says so. Compare the costs of 100BaseSX versus 100BaseFX,
> for example (which is much greater than 30 bucks).
>
> However, these are all points that would be addressed, and no doubt
> resolved, AFTER the general concept of 100Mbps P2P is approved. I am in
> favor of the general concept, and believe it meets the 5 criteria using the
> same arguments as the current set of P2P and P2MP. Architecturally, active
> topologies were left out of EFM for some reason. They are a natural
> complement to the current set of supported architectures, and should be
> included.
>
> --Dave Horne
>
> -----Original Message-----
> From: Ulf Jönsson F (ERA) [mailto:Ulf.F.Jonsson@era.ericsson.se]
> Sent: Tuesday, November 20, 2001 5:54 AM
> To: stds-802-3-efm@ieee.org; stds-802-3-efm-p2p@majordomo.ieee.org
> Subject: [EFM][EFM-P2P] 100Mbps P2P over SMF
>
> Hi all,
>
> On the January meeting we will make a presentation on the 5 criteria for
> 100Mbps P2P over SMF. Many of you have already indicated your interest and
> in order to have a broad acceptance of the proposal we invite all of you who
> would like to support, contribute or participate in this presentation.
>
> Please respond to me if you are interested. We intend to have a first draft
> presentation during next week that we can discuss off-line.
>
> Best regards,
>
> Ulf Jönsson
> Ericsson
> +46 70 2673313
--
Dora van Veen, Ph.D.
\ / Bell Laboratories/Lucent Technologies
((^\ -= * =- Broadband Access Research Department
\\ \() / \ Murray Hill, NJ, USA
>>/(_.'
((/\\
/|