In my May 12 article Powerline Communications Standards Continue to Struggle, I questioned the need for another wired home-networking standard like the ITU’s forthcoming G.hn standard. I’ll admit that the lack of details about this effort led to my doubt and somewhat negative reaction.
Feeling uncomfortable, I went back to the HomeGrid Forum for an update and a more in-depth understanding of the rationale for this standard. I recently spoke with Matt Theall, president of the Forum, and Kim Lewis, the co-chair of the Forum. Both are Intel employees.
THE FRAGMENTATION ISSUE
There are many wired home-networking technologies vying for attention and market share. MOCA and HomePNA use existing home coax for interconnecting TV sets, set-top boxes (STBs), DVD players, PCs, and other electronic devices around the home. Both are widely deployed, MOCA in Verizon’s FiOS fiber network and HPNA in AT&T’s U-verse IPTV network, but are not ubiquitous.
Then there are the powerline home-networking guys including HomePlug, HD-PLC, UPA, and IEEE’s P1901. HomePlug has the most product in the field, but none of these technologies has a solid grip on the market.
The main problem is the incompatibility of the technologies. Not one will interoperate with any of the others. And in most cases, the different technologies step on one another during operation, meaning they cannot co-exist without interference.
This is a nightmare for consumers who are expecting some level of compatibility, interoperability, or at least co-existence. Yet they do not exist. The resulting confusion has prevented the widespread deployment of these wired technologies.
Today, wireless (Wi-Fi) dominates the home networking space and it will continue to be widely used. But wired technologies could play a larger role if the confusion and lack of interoperability disappeared, explaining the G.hn effort.
G.hn is short for the ITU-T’s G.9960 standard effort. This worldwide project plans to generate a wired home-networking standard that works for most technologies. The ITU’s objective is to create a single next-generation physical-layer/media-access control (PHY/MAC) technology that runs over coax, phone lines, and powerlines.
The G.hn working group was founded in 2006, and major accomplishments have been achieved on the way to a final standard. The PHY and general architecture was given consent (ITU’s terminology) late in 2008 so the silicon vendors could begin their product work.
Currently the data link layer (DLL) has progressed to Baseline Text, a step before final consent. The PHY is generic orthogonal frequency-division multiplexing (OFDM) over the 2- to 30-MHz range that should work over any wired medium. Some of its key features include:
• A time-division multiple access (TDMA) media access control (MAC), which provides guaranteed bandwidth and low latency for applications with strict quality of service (QoS) requirements, such as IPTV. G.hn's MAC also supports multiple priority levels to ensure that an application, such as peer-to-peer file transfers, does not impact the performance of higher-priority applications like IPTV, VoIP, or online gaming.
• A logical link control (LLC) sub-layer guarantees delivery of data over home electrical wiring. G.hn’s LLC employs an advanced selective automatic retransmission request (ARQ) protocol that automatically re-transmits data affected by noise, and provides error-free end-to-end Ethernet services to G.hn devices on the network connected to powerlines, phone lines, or coaxial cables.
• A procedure for efficient aggregation of multiple Ethernet messages over a single MAC protocol data unit (MPDU). This significantly increases network throughput to support applications, such as high-definition IPTV, that have large bandwidth requirements. G.hn specifies line data rates up to 1 Gbit/s, much higher than current generation wired-networking technologies.
• State-of-the-art security based on AES (Advanced Encryption Standard - FIPS-PUB-197) and CCM (Cipher Block Chaining Message Authentication Code - NIST-SP-900-38C).
It is expected that the full G.hn specification will be given consent later in 2009 with first silicon available in early 2010. End products should follow later in 2010.
The HomeGrid Forum was formed to support and promote the new standard. Future testing and interoperability certification of products are intended to occur in the future. The Forum has heavyweights on its board, including Infineon, Intel, Panasonic, and Best Buy, with British Telecom (BT) as a recent addition. A host of other chip and end-product companies provide full support.
Silicon vendors like CopperGate, Ikanos, and DS2 are already working on G.hn chips, with Aware working on the software. Others are expected to join this effort. Companies now making powerline and coax-technology chips may join the G.hn effort to get a piece of the pie.
It is unclear whether they will offer dual-mode chips that support G.hn as well as their own technologies, which is doable but expensive. Once everyone is on board with G.hn, hopefully we can look forward to a wired home-networking approach that will be easy to use and fully interoperable, or at least with co-existence of all future products.
BUT WAIT. THERE ARE OTHER SCENARIOS.
While G.hn remains a dream waiting to come true in the marketplace, the other wired technologies are still selling. In the powerline arena, Home Plug has the dominant share. What will become of HomePlug when G.hn emerges?
I spoke recently with Mike Wilson and David Sorensen of Gigle, a fabless semiconductor company getting ready to announce new lower-cost HomePlug chips. Up until now there has only been one semiconductor company making the Home Plug chips, Intellon. Gigle thinks that a sole chip source has hurt the adoption of that standard.
Now with Gigle’s contribution, HomePlug could surge. That technology already has potential with a head start and the greatest market share. There is also a rumor that STMicroelectronics will make HomePlug chips.
Furthermore, Home Plug is working on its next-generation technology, which will further boost data speeds on the power lines. Sounds like the HomePlug option is still viable, but it remains to be seen who will be the ultimate winners in this classical-standards and technology battle.