Cable Ethernet allows remote segments of a WAN to communicate as if they were all linked by Ethernet cables. Of course, you can’t just string a 100Base-T cable across 50 miles. That’s where the “carrier” part comes in. Something has to get the Ethernet packets between the two locations. The advances in DOCSIS 3.1 make Ethernet over HFC an attractive option. The next step — DOCSIS 3.1 Full Duplex — will expand the possibilities even more?
Ethernet Service Distribution is still a big potential revenue source. But what is the role of orchestration in providing automation for Ethernet services? Learn more in our blog
Ethernet Service Distribution Through Orchestration.
Why Ethernet over DOCSIS?
Understanding the combination of Ethernet and DOCSIS takes a bit of explaining. They’re both protocols on Layers 1 and 2 of the OSI model, so why would you use both together?
The reason for wanting Ethernet is the efficiency of low-level protocols. Remote portions of a WAN can use TCP/IP over DOCSIS, but it imposes a high level of overhead. Direct Ethernet communication is more efficient. One or more Ethernet Virtual Connections (EVC) can be set up between two remote locations. They can be point-to-point or multipoint.
The difficulty is getting Ethernet connections over a long distance. The original idea was a protocol designed for a LAN. Pure Ethernet doesn’t scale well when the endpoints are miles apart. Dedicated carrier Ethernet approaches, such as MPLS, are expensive and lack flexibility.
Ethernet over DOCSIS, using existing cable networks, is more economical. Carrying Ethernet packets on DOCSIS adds some overhead, but it’s less than TCP/IP. DOCSIS includes security features to prevent interception or alteration of Ethernet data. The combination is an efficient and cheap way to bring locations together in an SD-WAN.
In the past, DOCSIS speeds have been slow compared to pure Ethernet speeds. DOCSIS 3.1, released in 2013, closes the gap. It allows up to 10 Gbps downstream and 1 Gbps upstream. The earlier DOCSIS 3.0 provided no more than 42.88 Mbps downstream and 10.24 Mbps upstream.
This doesn’t always mean that a connection can get gigabit speeds. A cable infrastructure designed for version 3.0 may not have a good enough signal-to-noise ratio to take advantage of the higher speeds. But a network can upgrade to the necessary cables and switches. This way it can provide customers with speeds at the top of the specification.
The benefits of speed-hungry operations such as database access, continuous backup, video streaming, and conferencing are obvious.
Full-duplex DOCSIS 3.1
There’s one drawback to DOCSIS 3.1: the asymmetry between upstream and downstream speeds. For many Internet use cases, the difference is reasonable. Usually, you download more than you upload. Within a WAN, though, “upstream” and “downstream” have no meaning. You want to speed in both directions.
The answer to this is still in development but coming soon: Full-duplex DOCSIS 3.1. With this, endpoints will use the same spectrum and speed for simultaneous connections in both directions. Not only is upstream communication as fast as downstream, but both can also happen at the same time. In previous versions, the available bandwidth splits between upstream and downstream data. One has to wait for the other. A proof-of-concept test by Nokia achieved 10 Gbps in both directions over a 100-meter cable.
The full-duplex enhancement eliminates the 1 Gbps bottleneck. This way it allows WANs to have the fastest connection over DOCSIS. It will be backward compatible with previous versions, though it will require new modems to take advantage of it. They didn’t announce an availability date yet.
Residential Bandwidth is growing with new access technologies for Cable and Telco companies. Cable has advantaged the DOCSIS 3.0 but Telcos are answering with PON/FTTH.
Keep reading here!