pon's and ethernet to the home
Hi, I've been reading a little about passive optic networks and the idea is very good from my stand point. As far as I have understood, the idea is to use the fiber as it was coax, doing some kind of FDM (frequency division multiplexing) with the lambdas (somehow the same). This would give us the capability to move at leat n x 10mbps ethernet on the same fiber using diferent lambdas for each customer, until power budget goes down. If the idea is correct, this would mean "next jump" on bandwidth. Who would be making this "ethernet/lambda multiplexors" right now? Is it feasible to do it today? or should we wait a little more? I mean, there are solutions using packet over sonet or alike, but pure ethernet? -- Miguel Mata-Cardona Intercom El Salvador mmata@intercom.com.sv voz: ++(503) 278-5068 fax: ++(503) 265-7024
On Tue, 9 Dec 2003, Miguel Mata-Cardona wrote:
As far as I have understood, the idea is to use the fiber as it was coax, doing some kind of FDM (frequency division multiplexing) with the lambdas (somehow the same). This would give us the capability to move at leat n x 10mbps ethernet on the same fiber using diferent lambdas for each customer, until power budget goes down.
The units that I worked with in the past (from an outfit called Quantum Bridge www.quantumbridge.com) worked pretty well, but the deployment model didn't meet my emplolyer's biz model (since it required outside plant). Basically the way it worked was it ran ATM OC-3 (mabe OC-12) down the fiber, using passive optical splitters, to the CPE. The CPE used a TDM style muxing for the return on a different lambda on the same fibre. The CPE itself was interesting too. It provided a 100bt ethernet port (showed up as and ATM pvc at the headend, with the actually allocated bandwidth controllable from 1.5 to 100 megs, as well as 4 DS1 ports. The DS1s were Circuit Emulated over the ATM fabric. Concerptually it was very interesting, and I imagine other POS solutions are similar, as simply using a lambda per customer would not be an efficient utilization of the available bandwidth. -Scott --- Scott Call Router Geek, ATGi, home of $6.95 Prime Rib I make the world a better place, I boycott Wal-Mart
www.carrieraccess.com makes PON CPE gear. http://www.carrieraccess.com/products/index.cfm/fuseaction/ default_prod/cat_id/118.htm www.alcatel.com makes PON 'head end' gear that works with CAC CPE. Basically, 1 strand of fiber (not a pair) can be used for 16 or 32 customers and will handle up/down data, down video, up/down T1 for voice at the customer. Head end voice, video and data is split apart. Carrier Access Corp hardware is rock solid, I have *never* had one fail. I don't use the PON stuff but I do use their DS1 & DS3 stuff. -Matt On Dec 9, 2003, at 12:58 PM, Miguel Mata-Cardona wrote:
Hi, I've been reading a little about passive optic networks and the idea is very good from my stand point.
As far as I have understood, the idea is to use the fiber as it was coax, doing some kind of FDM (frequency division multiplexing) with the lambdas (somehow the same). This would give us the capability to move at leat n x 10mbps ethernet on the same fiber using diferent lambdas for each customer, until power budget goes down.
If the idea is correct, this would mean "next jump" on bandwidth. Who would be making this "ethernet/lambda multiplexors" right now? Is it feasible to do it today? or should we wait a little more? I mean, there are solutions using packet over sonet or alike, but pure ethernet?
-- Miguel Mata-Cardona Intercom El Salvador mmata@intercom.com.sv voz: ++(503) 278-5068 fax: ++(503) 265-7024
-- Matthew S. Crocker Crocker Communications, Inc. Vice President PO BOX 710 Greenfield, MA 01302 P: 413-746-2760 F: 413-746-3704 W: http://www.crocker.com E: matthew@crocker.com
As far as I have understood, the idea is to use the fiber as it was coax, doing some kind of FDM (frequency division multiplexing) with the lambdas (somehow the same). This would give us the capability
In the optics field it is usually called WDM (Wavelenght Division Multiplexing).
to move at leat n x 10mbps ethernet on the same fiber using diferent lambdas for each customer, until power budget goes down.
Nope; PON works by sharing the same channel among the users. There is a WDM use of the fiber only because the uplink and downlink use different wavelenghts, but the downlink is shared with TDM (Time Division Multiplexing) and the uplink with TDMA (TDM/Multiple Access).
If the idea is correct, this would mean "next jump" on bandwidth. Who would be making this "ethernet/lambda multiplexors" right now? Is it feasible to do it today? or should we wait a little more?
PONs are feasible today; the CO equipment is called OLT (Optical Line Terminator), and the CPE is a ONT/ONU (Optical Network Terminator/Unit). The only external device is a passive star coupler. Check out www.ponforum.org, www.fsanweb.org and www.efma.org
I mean, there are solutions using packet over sonet or alike, but pure ethernet?
PON has three flavors: APON/BPON, EPON (IEEE 802.3ah) and GPON. APON uses ATM framing, EPON uses Ethernet framing and GPON uses GFP (Generic Framing Protocol). GFP can also be used on top of SONET to make Ethernet-over-SONET, Rubens
Dear Miguel; Having heard no answer, I will take a shot : I actually think that EPONs have a good chance to be the future method of distributing video from the "cable" provider to the home. As they are passive, it minimizes the amount of equipment out there. A 10-Gigabit Ethernet running multicast IP (probably with some form of packet tagging like MPLS) could more than support all of the video and data needs of a typical cable head-end customer base. There is a PON forum http://www.ponforum.org/presentations/page115.html and also the fiber to the home council which seems to be hot on EPONs http://www.ftthcouncil.org/ You might look at alloptic as a equipment provider here http://www.alloptic.com/ On Tuesday, December 9, 2003, at 12:58 PM, Miguel Mata-Cardona wrote:
Hi, I've been reading a little about passive optic networks and the idea is very good from my stand point.
As far as I have understood, the idea is to use the fiber as it was coax, doing some kind of FDM (frequency division multiplexing) with the lambdas (somehow the same). This would give us the capability to move at leat n x 10mbps ethernet on the same fiber using diferent lambdas for each customer, until power budget goes down.
If the idea is correct, this would mean "next jump" on bandwidth. Who would be making this "ethernet/lambda multiplexors" right now? Is it feasible to do it today? or should we wait a little more? I mean, there are solutions using packet over sonet or alike, but pure ethernet?
-- Miguel Mata-Cardona Intercom El Salvador mmata@intercom.com.sv voz: ++(503) 278-5068 fax: ++(503) 265-7024
Regards Marshall Eubanks T.M. Eubanks e-mail : tme@multicasttech.com http://www.multicasttech.com Test your network for multicast : http://www.multicasttech.com/mt/ Our New Video Service is in Beta testing http://www.americafree.tv
Having heard no answer, I will take a shot :
I actually think that EPONs have a good chance to be the future method of distributing video from the "cable" provider to the home. As they are passive, it minimizes the amount of equipment out there. A
May be not... all xPON systems have a wavelength window reserved for DWDM broadcast, that seems more suited to the task.
10-Gigabit Ethernet running multicast IP (probably with some form of packet tagging like MPLS) could more than support all of the video and data needs of a typical cable head-end customer base.
EPON/IEEE 802.3ah is a 1-Gigabit system (actually a 1.25 Gbps but 20% is used by the 8B10B encoding); I haven't seen any spec for a 10-Gig PON; even recently standardized GPON is limited to 2.5 Gbps downlink, 1.25 Gbps uplink.
You might look at alloptic as a equipment provider here http://www.alloptic.com/
Or http://www.alcatel.com/fttu http://www.flexlight-networks.com/ http://www.broadlight.com http://www.teknovus.com Rubens
participants (5)
-
Marshall Eubanks -
Matthew Crocker -
Miguel Mata-Cardona -
Rubens Kuhl Jr. -
Scott Call