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

Re: [EFM] EFM Requirements

It seems to me though that we have missed looking at a very viable 
If the PON we are looking at is 16:1 split we are really talking about a 
neighborhood splitter.
If that is the case, then 1 Gig p2p fiber to an active switch driving 
customer copper (HFCu) seems very viable to me given the following sort of 
	* We seem to be coming to a requirement for an active device at the 
customer demarc for isolation plus OA&M.
	* A "neighborhood switch" could have reasonable lengths for the copper.
	* We could backfeed the power upstream to the switch via a phantom circuit 
from the demarc box.

This would restore us to a network topology based on point-to-point links. 
This is the configuration that originally got Ethernet to "take off" due to 
simplified troubleshooting and failure isolation. If you look back to 
coax-based systems for Ethernet, the thing that was hated about them was 
doing the multi-point troubleshooting.

Regarding earlier comments on the reflector regarding powered devices in 
the field.

It has been noted that there are a lot of active/powered devices in the 
field for traditional bandwidth systems to providers. Every time the 
decision was made to put power in those I am sure that there was a lot of 
wishful thinking that they could be deployed without a power supply. This 
wish was eventually set aside when it was discovered that there is no such 
thing as a free lunch. There are powered devices widely deployed in the 
field because the cost/performance/maintenance trade-offs put them there.

| Geoffrey O. Thompson                    |
| Nortel Networks, Inc.  M/S SC5-02       |
| 4401 Great America Parkway              |
| P. O. Box 58185                         |
| Santa Clara, CA 95052-8185  USA         |
| Phone: +1 408 495 1339                  |
| Fax:   +1 408 495 5615                  |
| E-Mail: gthompso@xxxxxxxxxxxxxxxxxx     |