There is a growing interest among Telco/Cable Service Providers in deploying networks that can take advantage of all the existing wires available in the home (power lines, phone lines and coaxial cables).
Until recently, there was a number a proprietary technologies for wired networking, but each one of them was designed for one single medium (ie, one technology for powerline only, another for coaxial cable only, and so on).
Each of these technologies is fundamentally different from each other, so networking chips have to be specialized: For example, a chip designed to operate over coaxial cable cannot work over powerlines.
Why are the chips different? In many cases, the reason is simply that they were designed in separate groups, inside closed organizations, which took different decisions when choosing technical elements such modulation type, Forward Error Correction, QoS architecture, Security architecture, etc.
The end result is that a device (such as a Set Top Box or a Residential Gateway) that wants to support both coax and powerline networking using these propietary technologies needs two completely different chips.
This option is far from optimal: having two chips means twice the cost, twice the power consumption and twice the board space. These disadvantages explain why there are few products (or none at all, as far as I know) that implement coaxial networking specs and powerline networking specs simultaneously. Although some vendors are desperately promoting solutions that require multiple chips, in my opinion these are not cost effective, and are not adequate for the price-sensitive Service Provider market.
Also, these technologies cannot be easily integrated in a single chip, as they use different modulation schemes, different FEC schemes, different QoS and security architectures, etc. Integrating two of these technologies in a single chip would make a chip twice as big.
Addressing this problem was one of the main reasons for the development of G.hn in ITU. G.hn specifies a single PHY/MAC that can operate over any wire. This means that silicon vendors can design one single chip that can work over power lines, phone lines and coaxial cable. Equipment vendors only need to embed one chip that can do it all. The end result is a solution with half the cost, half the consumption and half the board space as alternative solutions with multiple non-interoperable chips.
Additionally, silicon vendors that develop G.hn chips and equipment vendors that manufacture G.hn devices can address a single large-volume unified market (as opposed to multiple smaller fragmented markets), providing further advantages in terms of economies of scale.
Recommendation G.9960, which specifies the Physical Layer of G.hn, received final approval at an ITU meeting in October 9th 2009. G.hn has already garnered support from multiple silicon vendors, service providers and equipment manufacturers. G.hn silicon is expected during the first half of 2010, with end-user products shortly thereafter.
Comments