How We See CE

In the year since my original post to this blog, "Setting the standard in wireline home networking", our industry has devoted many engineer-hours to progressing the wireline home networking standards currently in development. Now that we can see the light at the end of the tunnel, it is time to turn our attention from setting the standard to delivering it.

First: an administrative point. In this post, I represent my own thoughts and opinions. I am not representing the views of the industry groups and standards groups mentioned. Next: a recap. The wireline home networking industry wisely concluded a few years back that a SDO (standards development organisation)-endorsed standard was a necessary prerequisite to the mass market, and that proprietary solutions were condemned. A project was established in IEEE to develop a standard for in-home and broadband access powerline communications. A little later, a question was established in ITU-T to develop a Recommendation for home networking over coaxial cable, powerline and phone lines. Now fast forward three years.

The IEEE P1901 working group sets a high bar for decision making. A 75% majority of participating entities is required to make any significant decision. The original composition of the group often meant that stalemate was inevitable. Then the Homeplug Powerline Alliance (of which my company, Gigle Semiconductor, is a board member) and Panasonic decided to merge proposals to move forward. Around a year ago, when arbitrary entities with no obvious knowledge or interest in powerline communications started to attend meetings with the intent of effecting decision making, the IEEE hierarchy stepped in and introduced a revised and fairer set of rules requiring meeting participants to understand the project, be competent enough to make technical judgements, and regularly participate. As a result the engineers are back, the marketeers, politicians and expensive standards-consultants have gone, and the arbitrary entities are no longer eligible to attend. Copies of Roberts Rules of Order, previously compulsory reading, are conspicuous by their absence. The net impact has been a year of positive discussion, refining and maturing the standard, with all substantive decisions being passed by the group supermajority with a like mind.

In December of last year, the group voted to approve the IEEE 1901 Baseline Standard. At this time the working group, looking at activities taking place in ITU-T, introduced a placeholder for a potential 3rd "G.hn compatible" {HY (physical layer) in order to allow for its inclusion in 1901, in the event that the G.hn activity incorporated capabilities for access applications. In the opinion of some, however, this augmentation rendered the standard somewhat irrelevant, and it was subject to justified criticism as a "3 PHY monster". In July, the Draft Standard was approved. At the subsequent meeting this October in Boston, the group’s technical experts concluded that the G.hn activity did not incorporate the required access requirements, and the group voted to remove the placeholder for the third PHY. In addition to that, the group made substantive progress in resolving comments received on the Draft Standard, and is in very good shape to move forward to the final steps of Sponsor Ballot and (finally) publication. Suddenly, IEEE 1901 is back on the map again. Two PHYs works (for example: IEEE 802.11 a, b). Three is a crowd.

Meanwhile over at ITU-T, Study Group 15 Question 4 (Q4) has been progressing on its G.hn work. Here, decisions are made by consensus and while this can, on occasion, be a lower threshold than the 75% required by IEEE, it does have the benefit of allowing pragmatic decisions to be taken in the best interests of group progress. The G.hn companion industry group, the HomeGrid Forum (Gigle is a Promoter Member), has evolved considerably over the last year, adding Best Buy, British Telecom, Sigma Designs and Coppergate to the Board of Directors. In addition, AT&T has lent its muscle to the cause with vocal support and participation in joint webinars. It has been widely reported that Q4 had been targeting to achieve consent of the complete G.hn specification this year, but it did not make sufficient progress. The group did approve the PHY that had been consented in December 2008, but the MAC, the Data Link Layer and an amendment to the PHY were not consented, rather they are placed in "Q4 determination"; in layman’s terms, this means 'not cast in stone but can only be changed to correct a technical error".

While all this has been going on, the SIGs (special interest groups) are not standing still. MoCA is creating MoCA2, which aims to secure its position as the defacto standard in RF coax in North America and take it further afield. HomePlug is creating HomePlug AV2 which targets higher levels of coverage and throughput on powerline, and HomePlug GP (“green PHY”) for smart grid applications; both new initiatives are fully interoperable with Homeplug AV.

At the same time, the market has seen two major structural changes. The first is the entry of top-tier silicon vendors into the powerline communications space. With Atheros’ acquisition of Intellon and its implied admission that WiFi on its own is not enough for home entertainment networking, the phrase of the moment is "HomePlug is the new WiFi." STMicroelectronics has also announced its intent to enter the HomePlug/IEEE 1901 market. Secondly, and equally important, is that HomePlug AV has become a multivendor standard with certified silicon available from two vendors (Intellon/Atheros and Gigle) and with others in the pipeline. HomePlug is currently holding its first HomePlug AV / IEEE 1901 silicon interoperability plugfest. These evolutions considerably reduce supply chain risk and will help to deliver the standard...

...which brings me to the main point of this post. Setting a standard is one thing. Delivering it is quite another. A standard only exists when there are at least two certified product vendors. Otherwise, it is simply an interesting piece of paper. In the rush to applaud the consent of G.hn and the approval of the IEEE 1901 Baseline last December, we had some rather optimistic announcements of silicon availability from some. The claims of first silicon by the end of 2009 have proven to be somewhat unfounded. In any case, who cares about first silicon? It is an indeterminate milestone between a written specification and a certified interoperable product. The only milestones that truly matter are when a standard has a proven compliance and interoperability certification program at when there are at least two certified interoperable products. Watch this space.

As I predicted in my previous post, it looks like the current standards initiatives are likely to take us a step closer to the ultimate goal of a single unified standard, but we will need another consolidation step to finally get there. Today there are seven notable non-interoperable pseudo-standards for wireline home networking. When IEEE 1901 and ITU-T G.hn are complete we will have three – HomePlug / IEEE 1901; MoCA and G.hn. From a market adoption standpoint the options for silicon vendors are becoming clear. Powerline products will need to support several standards. HomePlug AV / IEEE 1901 and HomePlug GP will dominate retail, consumer and smart energy markets. G.hn combined with HomePlug AV will be required for those service providers who wish to deploy G.hn but avoid the service calls from customers complaining that their network suddenly doesn’t work when they add HomePlug AV products from retail. HomePlug AV2 will be required for those service providers who want to provide an interoperable path forward for their existing HomePlug AV customers.

Michael Wilson
Gigle Semiconductor