Recent sparks of interest in passive optical networking (PON) may justify a reborn faith in this nascent market, despite the impatient cries of naysayers.
PON equipment seeks to lower the cost of access networks by extending optical bandwidth without requiring more expensive "active" components such as lasers or amplifiers (see PONs: Passive Aggression for a primer).
So far, however, there are only about 30 deployments of PON technology worldwide, nearly all of them with small carriers and cable operators seeking to serve small businesses and residential subscribers.
Still, more than $527 million has been invested in private PON companies over the last two years -- at least half of all startups have received new funding this year.
These mixed messages have fueled a growing impatience with the slow realization of PON in carrier networks. While the technology continues to boil away on broadband’s back burner, many are questioning whether the still-tiny market will ever gain its legs.
I have some questions need help, I know that if EPON wants to compare with APON, it need to establish priority-based traffic control and providing traffic shaping. My questions are,
1. In additional to traffic shaping, do we need to have policing (e.g., dual-leaky bucket or 3-color marker) and Congestion control (e.g., WRED) installed on EPON ?
2. What I guess is that we need to perform Policing on upstream, and shaping on downstream. Is it correct ?
Hello Doco, I´m planning the design of a APON network over a existing SDH network. We want to reuse the existing infraestructure as much is possible. So I want to ask you if you know if the OLT can be located at the PoP (point of presence), that it means, in the node primary locations, an then connect the OLT with the SDH equipment, because they have this kind of interfaces. If you konw how to do this, please, explain to me, because I have a lot of doubts in this area. Thanks very much and, I expect your answer. Mª Jesus.
1MB with potential of higher speed. Isn't this a low-cost competitor? ____________________
One is lucky to get 1Mb/s in bursts. Most of the time its much less. Also, the upstream rate limiting prohibits services like backing up the family photo server to a colo locker box.
The following download times may help one understand the potentials. (Note: Realtime for most human senses require response times less than 3ms) First the sizes
A) Photo (500KB) B) Song (4MB) C) Movie (3GB) D) Library of Congress (50TB)
And the download times (no congestion assumed)
A) 768Kbs DSL - 5.2s, 42s, 8.7hrs, 17 yrs B) 100Mbs FE - 40ms, 320ms, 4min, 47 days C) 10Gbs - 4us, 3.2ms, 2.4s, 11hrs D) UPS - days, days, days, flight to DC
Network owners building infrastrucutre to support discount consumptions should look to Walmart for their business model. Imagine if Walmart charged access fees and built parking lots for horse and carriages. We'd laugh at them.
The article compares high speed DSL with PON. Seems like cable modem offers very high speed, 1MB with potential of higher speed. Isn't this a low-cost competitor?
> Would the token bus architecture work, > creating a logical ring and > passing around an access token?
No - not readily. The head end can receive and transmit to each of the ONUs. The end points (ONUs) can receive from and transmit to the head-end. BUT an endpoint can not directly send a message to another endpoint.
Because of this, the endpoint would need to tell the head end when he is done transmitting, and then the head-end could tell the next node to start. Because of the latency of sending a signal 20km, this would end up with large chunks of time not being used.
You also have a tradeoff in latency vs. % of bandwidth available. With the larger guard bands it is more efficient to send larger bursts of data at a time. However, then by the time each of 32 nodes sends a large burst, and you have all the guardbands, you might be talking an additional latency of 30ms or much more. Adding that much latency would put you into the range where you would need to add echo cancelation for voice traffic. Echo cancelation is expensive, so if it can be avoided by having some extra complexity by doing the ranging process it ends up being a cheaper system.
The tricky part in both of these is that either a very large guard band of time is needed between each burst of traffic, OR every end node needs to be told how far they are from the headend. ______________________
Would the token bus architecture work, creating a logical ring and passing around an access token?
I have been hearing about the PON and HFC technology for so many years. In a way it is similar to hearing about MMDS and LMDS.
QWEST has been caught in its own game of wasting company resources. Unlike any other RBOC it has lost 70% of its markwet value.It had dedicated a significant amount of resources for testing all products under the sun. Against the conventional wisdom, QWEST is also deploying Softswitches. This will not let buy anything new for QWEST customers except unstability.
SBC has aklso has lost about 40% of itsmarket value during last year. The company is now the largest RBOC and is not in a position to deploy new technologies unless it becomes profitable by cutting back on its workforce. SBC also wastes a lot of its resources on trying things that are not needed.
Many other carriers are not likely to explore PON technology for deployment.
The blogs and comments are the opinions only of the writers and do not reflect the views of Light Reading. They are no substitute for your own research and should not be relied upon for trading or any other purpose.
To save this item to your list of favorite Light Reading content so you can find it later in your Profile page, click the "Save It" button next to the item.
If you found this interesting or useful, please use the links to the services below to share it with other readers. You will need a free account with each service to share an item via that service.
Digg
Del.icio.us
Reddit
