For many years questions about using 75ohm coax for ethernet regularly appeared here. After a while, they disappeared.
It seems now that the devices are being made, using either the MoCA or HPNA protocols.
The primary use seems to be rewiring houses for IP based digital television. One can add a DSL modem/router/access point including a coax output, connect to the existing home coax cabling, then put set-top boxes that accept IP over coax and output the appropriate video signal.
It seems that bridges are available, though somewhat expensive, to bridge between standard ethernet and coax.
There seems to be the HCB1000 for ethernet to HPNA, and ECB2200 for ethernet to MoCA.
I noticed the same thing. A big user of this is Verizon, in their FiOS system. They use MoCA. It makes a lot of sense. You can reuse the same coax cabling that the cable companies installed, passive spitters and all.
But it's not the Ethernet we're accustomed to. It's more a case of sending Ethernet frames down a passive coax cable plant, and upstream as well, and all depends on the active head-end. There is no carrier sense/collision detection or Ethernet layer 2 switching, when it comes to what is actually going on in the 75 ohm coax parts of the network. The active head-end takes care of that. Downsteam and upstream traffic is frequency-divided and time-divided by the one traffic cop, the active head-end. What would otherwise be a distributed Ethernet choreography is instead strictly controlled by the head-end.
So for example, if you install just one of these cable modems in your house, then you would have to go to Cat 3 cabling in a hub architecture to reach the other IP devices in the house. Or WiFi. So, none of that in-home networking is over 75 ohm coax anymore.
If you install more than one of these modems in your house, to connect to several PCs or TVs, using the passive splitters in the coax plant to reach multiple rooms, then all of the traffic goes between the "head-end" and each of these in-home appliances. So for example, for two PCs in your home to communicate, they have to go way upstream to the head-end first, then back down. The traffic doesn't just go to the closest passive splitter, then back to the other in-home PC, as one would expect Ethernet to do.
Not to detract from the scheme. Just saying that I don't consider to be Ethernet in the traditional sense. Just something that hauls Ethernet frames around, using its own traffic control scheme.
As far as I can tell, you can use two of the same boxes, or more, for more connection points. They must somehow vote for a head-end then.
Also, wouldn't it still be CSMA/CD for access to the upstream data stream?
I didn't read the standard, but only the specifications page for some of the boxes.
It may go to the head end, but it should also come back to all the other boxes. As far as I know, the usual use is coax for TVs, and UTP for computers, but it doesn't have to be that way. There seem to be PCI cards, too.
They make it sound like full-duplex, but it might be that they just put enough bandwidth on the system such that two ends can talk at 100baseX speeds without any collisions.
Well, do remember that ethernet includes 10broad36, which also goes to the head end. So that IS part of ethernet. I don't know if either HPNA or MoCA is really similar to 10broad36, though.
The FiOS system is fiber to the home. Then, inside the home, as far as I can tell, they create what amounts to (functionally) a mini-DOCSIS cable plant. If you have multiple modems inside the house, to distribute the service over the passive 75 ohm coax plant, you would have one appliance attached to each modem. All traffic to and from each of these modems goes to their in-home gateway, and then from that gateway to the main box where the fiber cable is connected from outside. So that gateway is the traffic cop.
Don't think so. Each modem gets its turn, round robin.
Good point. I keep forgetting about 10BROAD36. The main differences are that 10BROAD36 does have a collision detection scheme, it is purely half duplex, and is limited to 10 Mb/s of data carrying capacity over a 6 MHz channel. MoCA uses 50 MHz channels. DOCSIS uses
6 Mhz channel(s) for downstream.
But your point is good. Even the first full duplex switched Ethernets didn't appear to be "true Ethernets," because we hadn't seen any Ethernet that didn't rely on CSMA/CD before. The new switches were much like ATM switches, only they happened to use Ethernet frames instead of ATM cells. So one could argue, what is Ethernet in these, other than the basic frame format?
"Synchronous MAX protocol controlled by a dynamically elected master employing a collision avoidance media access strategy."
It seems to be an extension from G.9951 and G.9952 which use an asynchronous MAC protocol.
As far as I can tell, for HPNA 3.1, the data doesn't go to the head end and back (10broad36 style), but the head end (master) does schedule transmission. If there is a RG (residential gateway, previously known as modem/router), then it is the master, otherwise the stations select one to be the master.
Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.