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

[RE] Residential Ethernet services as part of 802.3

Title: Residential Ethernet services as part of 802.3
There are a number of ways that the ResE objectives can be met using both 802.1 and 802.3 processes, but since we are an 802.3 SG, it might be best to organize things so that we could do the most work as part of 802.3. Here is a possible approach (this makes more sense if you are familiar with the presentation I made in Ottawa on technical and economic feasibility, you can get it at http://grouper.ieee.org/groups/802/3/re_study/public/sep04/teener_2_0904.pdf):

  1. ResE devices are full duplex, so we can think about all ResE connections as between two peer stations. (Bridges appear to be N ResE stations if they have N ports).
  2. Following the style of several recent efforts, we use the MAC control layer (sublayer?) as the way the new services are specified.
  3. A MAC control sublayer would be used for doing clock synchronization. Associated MAC control PDUs would be exchanged to keep the isochronous service timer on each peer station synchronized. Associated MAC control services allow higher layers provide “directionality” to the synchronization (which peer follows the other). Default could just be highest station number. (A future 802.1 update would use this service interface to keep a bridge and the whole bridged network synchronized to a single station). Timing services provided to higher layers would be used for time-stamping packets to be transmitted, and delaying received packets until their “presentation time” matches (see IEC 61883 for how this procedure is done for compressed and uncompressed audio and video in consumer electronics). The isochronous packet service sublayer (see below) would use these services to determine whether queued best effort or isochronous packets should be transmitted .
  4. Another MAC control sublayer would be used to communicate admission control requests (bandwidth requests) from a receiving device back to the peer transmitter and how the transmitter would use those requests/reservations. (A future 802.1 update would describe how these requests are forwarded in a bridge).
  5. Finally, there would be a isochronous packet service sublayer that would determine when best effort or isochronous packets would be launched. Isochronous packets would need unique labels (perhaps a unique length/type value) so they could be processed at the receiving isochronous packet service sublayer. (A future 802.1 update would describe how isochronous packets would be routed inside a bridge).

I can give a presentation describing all this in a lot more detail if it would be useful. It would take about 30-45 minutes. (I guess I’m asking for a slot, Steve.)
--
----------------------------------------------------------------
Michael D. Johas Teener - Mike@plumblinks.com PGP ID 0x3179D202
Plumblinks, 23 Acacia Way, Santa Cruz, CA 95062-1313
+1-831-247-9666, fax +1-831-480-5845
------------------- www.plumblinks.com ------------------------