In 2002, we ran a series of tests of HomePlug, a new (at the time) industry-defined standard for home networking over existing power lines. We were impressed by HomePlug's performance -- comparable with Wi-Fi at that time, and more consistent around our house.
As we were conducting these tests, we said to each other "We got it! Buy any device, plug it in, and it's networked!" Surely it wouldn't be long before every device with a power cord would be networked over the power line; Wi-Fi would be used only for portable battery-operated devices. Our article in September 2002 was headlined HomePlug Powerline Networking - Getting ready for prime time.
Eight years later, nearly all mobile devices that would benefit from networking -- laptop PCs, smart phones, iPads and more - come with Wi-Fi built in. By contrast, most fixed-location devices -- such as desktop PCs and the latest TVs -- still have only Ethernet. It's up to the unsuspecting buyer to figure our how to get them connected to the home network.
Wi-Fi has succeeded because the key players in wireless networking were able to agree very early on a single standard: IEEE 802.11, branded as Wi-Fi. The chip companies promoting other wireless approaches swallowed their pride and rallied around Wi-Fi well before most consumers realized they needed home networking. What was the payoff? A recent headline reads "Wi-Fi IC Shipments Could Top 770 Million in 2010".
Networking over existing wiring has underperformed because the players have never been able to agree on much of anything. Chip companies have lined up in opposing camps, each seemingly more focused on beating the other camp -- or their competitors in the same camp -- than on growing the pie to benefit everybody. Devices from one camp often interfere with those from another camp. Consumers can't tell them apart.
The result? Total IC shipments are only a tiny fraction of those for Wi-Fi. Very few consumer-purchased devices come equipped with any form of existing-wiring networking. One of the networking camps thought it was impressive to put out a press release proclaiming "Experts Agree: 10 Million ... Connected Homes by 2013."
This may be about to change. Most of the chip companies have aligned around two emerging standards: IEEE P1901 and ITU-T G.hn. These are very different standards, with no requirements for interoperability between them. But they are designed so as not to interfere with each other if both are installed in the home.
Chip makers say ICs for both standards will be on the market this year, with early devices shown at CES next January. Telephone companies - the primary supporters of G.hn - are poised to place orders for delivery next year. Consumer electronics companies may be getting ready to place bets on which standards to adopt for Internet-connected TVs and video media adapters.
This should all come to a head by the end of this year. We think next year will prove to be critical for the future of wired home networking.
Battle of the SDOs
Wired home networking has long suffered from fragmentation, with five different technology camps competing for dominance over existing electrical, coax, and phone wiring. Most of these technology camps have now lined up with either IEEE P1901 or ITU-T G.hn.
Until now, "standards" for networking over existing wiring came from industry consortia. By contrast, the IEEE and the ITU are long-established highly-respected standards development organizations (SDOs). Both have developed thousands of internationally-accepted formal standards.
Large service providers are moving toward wide-scale deployment of triple-play services. They view networking over existing wiring as far preferable to pulling new wires or using Wi-Fi. Both P1901 and G.hn are directed to large-scale purchases by service providers. The stakes are high, and all technology camps and their backers are gearing up.
At CES in January, we met with representatives of these camps and their key member companies. We recently followed up by phone to get an update on the current status and the outlook for 2011.
The IEEE P1901 Working Group brought together key members of two of the powerline camps (HomePlug and HD-PLC), along with many other industry representatives, to create a comprehensive standard for Broadband over Powerline (BPL). The P1901 standard includes detailed mechanisms for both BPL access to the home and BPL networking in the home. The camps were unable to agree on a single physical layer (PHY) approach, so P1901 includes two distinct PHY layers (one from each camp) with a common upper MAC layer.
Since the two PHY layers would interfere with each other if they were not coordinated, P1901 includes a coexistence mechanism that permits both PHYs to operate in the same house, sharing the potential bandwidth of the powerline wiring while gracefully deferring to each other. This coexistence protocol - Inter System Protocol (ISP) - is also explicitly designed to coexist with G.hn and with BPL used for broadband access to the home. P1901 envisions the possibility of four non-interoperable PHYs sharing the power wiring, with ISP keeping them out of each other's way.
At CES, we met with Rob Ranck of the HomePlug Alliance; we recently followed up on the phone. Much of our discussion concerned the progress of P1901 and its market impact.
The P1901 standard is very nearly finished. It passed its Sponsor Ballot in April 2010; the IEEE website says "Approval of P1901 as an IEEE standard is targeted for September 2010". The draft standard is available for purchase from the IEEE Website.
The HomePlug Alliance appears to be preparing to take on the role of certifying P1901 products. In April, HomePlug announced that it is working with HD-PLC on a "joint certification program" to test coexistence between the two PHY technologies within P1901, and said an interoperability "plugfest" had been scheduled for May 2010 "to certify the first compliant silicon chips."
ITU-T G.hn/HomeGrid Forum
ITU-T G.hn brought together key members of two other camps (HomePNA and UPA), along with other industry players, to work out an ambitious specification for an "every wire" standard designed to operate over all existing wiring: powerline, phoneline, and coax. Aiming at a higher performance level than existing technologies, G.hn did not adopt an existing protocol, but rather designed a new approach based on elements drawn from all the existing camps.
HomeGrid Forum, a trade group formed to promote G.hn, aspires to play the same role for G.hn that the Wi-Fi Alliance plays for IEEE 802.11. Its members include many companies active in the G.hn standards development. It is already actively promoting G.hn and is working to develop the certification testing process.
At CES, we met with a group representing HomeGrid, including Matt Theall of Intel, Mike Coop, John Egan of DS2, and Mario Finocchiaro of Aware. They said the G.hn standard was nearly done; indeed, the last pieces were consented in Geneva the following week and were approved in June 2010. At CES we also met separately with Michael Weissman, now VP Corporate Marketing at Sigma Designs. We met with Michael at the recent Cable Show, and talked on the phone with Matt.
Many silicon vendors are working on G.hn chips. DS2, Sigma and Lantiq are well along. DS2 and Sigma promise to have samples ready by the summer and fall. Matt says working interoperable devices should be shown at CES next year and that many more chips are "in development."
Earlier this week, HomeGrid and The Broadband Forum announced that they were collaborating on the global certification and interoperability program for G.hn.
MoCA -- the fifth technology camp -- is not supporting either P1901 or G.hn. Instead it is continuing on its own, and developing its own second-generation specification, MoCA 2.0.
We recently talked on the phone with Vinay Gokhale, SVP of Marketing and Business Development at Entropic Communications, the company that founded MoCA. Vinay described multi-room DVR as "the killer app". He said multi-room DVR has "really started in a big way" in North America -- and most video service providers use MoCA to connect set top boxes between rooms.
Vinay said that Broadcom's support for MoCA had played "an important role" -- it helped the video service providers "get comfortable" by having a major chip company as a second source. "Service providers take comfort from multi-source vendors; costs can be controlled."
We asked Vinay if he was concerned about competition from G.hn. He said G.hn was at least two years away from market. Nearly all major North American video providers have now committed to MoCA. MoCA has become "a self-fulfilling prophecy" -- it's "much more difficult to deploy without MoCA."
And MoCA 2.0 isn't far away. Earlier this week, MoCA announced the ratification of the 2.0 specification. It offers "Basic" and "Enhanced" modes that promise 2X and 4X the performance of MoCA 1.1, with full back-compatibility with the earlier 1.0 and 1.1. It includes low-power modes for energy efficiency, and higher-performance modes for point-to-point applications. Soon after the announcement, Entropic's CTO Tom Lookabough told us we should expect to see interoperable "MoCA 2.0 devices in volume by the beginning of 2012."
HomePlug "AV2" and "AV Turbo"
While P1901 and G.hn are moving toward the market, HomePlug and its member chip companies are not standing still.
While focusing much of its attention on the benefits of P1901, the HomePlug Alliance is continuing to work on the second generation of AV, called "AV2". When AV2 was announced, HomePlug said it would far exceed the performance of G.hn while maintaining back compatibility with today's AV. Rob Ranck recently told us that HomePlug was “insuring AV2 is true next-generation” technology.
Meanwhile, Atheros, Sigma Designs and Gigle have introduced chips that claim to perform much better than HomePlug AV while still maintaining back compatibility with AV devices.
In an announcement late last week, Sigma explained how they boost powerline performance. They said that their new HomePlug AV chips use MIMO (multiple input, multiple output) -- the same technique used in 802.11n Wi-Fi -- to get much better performance than competing chipsets. They said MIMO will also be used in their G.hn chips expected later this year.
The Outlook for 2011
We've written many times about what we term "whole home networking" -- integrated networking for digital voice, video and data. The need for data networking became clear as consumers needed to connect multiple PCs to the Internet. But the need for video networking was less obvious until IPTV, multi-room DVR, and Internet-connected TVs came along.
Now all video service providers - most also broadband service providers - see the need for whole home networking, and are trying to decide which technology(-ies) to deploy. Pulling new wires is time-consuming and expensive. Wi-Fi can't reliably and consistently handle multiple streams of high-definition video across an entire house. Service providers want to use existing wires to deliver digital data, video and soon telephony.
In North America, where nearly every home already has coaxial cabling between all the TV sets, it's an easy decision to use the same cables for digital video services. Verizon and most cable operators have chosen to use MoCA; AT&T and some smaller telcos have chosen HomePNA and say they expect to move to G.hn when it is ready. Many have used Wi-Fi and HomePlug for data networking.
Outside the US, many telcos have deployed Internet and IPTV services using either specialized Wi-Fi equipment or powerline with HomePlug AV, UPA, or HD-PLC. In markets where most homes already have coax, cable operators and telcos are likely to choose MoCA now or wait for G.hn. Many telephone companies participated in the ITU's G.hn effort, and have announced their intention to deploy it.
If the G.hn chip makers meet their projected schedules, they will show interoperable prototype devices based on two or three different chips by the end of this year. These early devices promise performance better than HomePlug AV and MoCA. Some will provide back-compatibility with HomePNA, UPA, or HomePlug AV, providing a migration path for the many telcos who have deployed these earlier technologies. If G.hn achieves these objectives, many leading telcos are likely to deploy G.hn devices for both IPTV and data networking. This would create a big market for G.hn chips and devices.
This opportunity has caught the attention of some big chip companies. Until recently, the existing-wiring networking business was the province of small early-stage fabless chip companies. Now some of the world's largest semiconductor companies sense a big opportunity for chips -- first in devices for telcos, then in all networked consumer electronics devices. Some of these big companies have lined up behind standards:
Several large chip makers are already Promoter members of HomeGrid Forum. If the telcos announce plans to place large orders, more chip makers will see the opportunity and commit to producing G.hn chips -- or SoCs incorporating G.hn along with other functionality. This will drive BoM costs down, lowering the price of retail devices, and providing a path for Internet-connected TVs and other video devices.
If the G.hn chip makers fall short, HomePlug chip makers will be delighted to show the telcos P1901 and "AV turbo" solutions which outperform AV while providing back-compatibility, and are deliverable now in quantity. MoCA will try to build on its North American success to capture market share where there's significant coax penetration.
By 2011, it should be clear whether G.hn will get the lion's share of the global existing-wiring market -- or whether P1901 and MoCA will be free to dominate the market.
What About the Consumer?
At every powerline display at CES 2010, we saw consumer electronics insiders surprised at the novel idea of using powerline to connect media devices. Many seemed to think there was a Wi-Fi antenna hidden somewhere inside. Powerline has been around a long time, but most people still don't get it.
While the Wi-Fi Alliance focused much of its early efforts on educating consumers on the benefits of using wireless for home networking, none of the existing-wiring camps has invested any significant effort to building consumer awareness. Our recent discussions with all the existing-wiring players revealed that their efforts continue to be focused on wooing the service providers and bashing each other. If existing-wiring networking is ever to achieve its true potential in the broadband home, they'll need to shift their attention to the consumer.
For More Information
We have written about home networking since the first issue of this newsletter more than ten years ago. Our Topical Index: Home Networking provides access to our articles on all aspects of home networking, organized by technology.
( www.homeplug.org ) ( www.homegridforum.org ) ( www.intel.com ) ( www.ds2.es ) ( www.aware.com ) ( www.sigmadesigns.com ) ( www.broadband-forum.org ) ( www.entropic.com ) ( www.broadcom.com ) ( www.atheros.com ) ( www.giglenetworks.com ) ( www.wi-fi.org )