Wednesday, June 17, 2009

Next-gen Wireless Home Networking Solutions: A Follow-up

Being the "home network analyst" at Parks Associates for 11 years, I've been involved in a few of these home networking PHY beauty contests that sometimes take on the form of high school hallway debates.

  • "The analyst likes so-and-so better than such-and-such!"
  • "He like so-and-so, but he doesn't "like-her/like-her!"
  • "The analyst took sides with so-and-so and says that such-and-such is [enter in derogatory statement here]!"

I've accepted that this comes with the territory. In the end, what I'm trying to do when I write a report, an article, or lead a moderated discussion among highly competitive players vying for pieces of the same pie is to put myself in the shoes of a home network vendor, a consumer electronics manufacturer, or a service provider and do my best to evaluate technology solutions as consistently as possible. This means that I get to ask a lot of "dumb" questions - often multiple times, and often over a multiple period of briefings over multiple years to the same companies.

To the home networking companies that have endured the questions and shared your perspective, thanks!

The stakes for next-generation wireless networking solutions are really high these days, as silicon companies target the embedded market well beyond the router, access point, and notebook computer space. Our recently-completed Home Networks for Consumer Electronics report forecasts that worldwide sales/shipments of network-connected (and in most cases Web-enabled) consumer electronics products will double between 2009 and 2010. That's a chipset opportunity well north of 100 million units in 2013, so it's no insignificant market. If you add in embedded home networking in devices like residential gateways, home computers, and pay TV set-top boxes, you're adding at least another 50-60 million units each year during the forecast period. So, the pressure is on, and with a reported $300 million in funding having gone into next-generation wireless chipset companies over years, folks want to make sure that they make the right calls.

We had a really great wireless networking panel at our CONNECTIONS™ Summit at CES in January (where panelists from AMIMON, Celeno, ProVision, TZero Technologies, and WirelessHD spoke), and the feedback was positive in terms of the information that was covered. TWICE provided an excellent summary of the panel. As we started planning for CONNECTIONS in June, we decided that it would be well worth it to bring some of the main players in next-generation wireless networking back for a reprise. Besides, I had plenty of dumb questions left to ask! So, we asked six companies to participate on a panel discussion titled Wireless Networking & Video Consumption in the Home.

First, I want to thank the participants - Vijay Desai from Aceurity, Noam Geri from AMIMON, Lior Weiss from Celeno, Ian Walsh from Provision Communications, Andrea Goldsmith from Quantenna Communications, and Sheung Li from SIBEAM - for competing with food and beverages in the Showcase Theater, and generating a lively debate about the respective merits of their solutions! Who knew that one simple question would lead us down the 75-minute road we took?

Several of our panelists have agreed to answer some follow-up questions, so I've captured their responses. My hope is to get some additional perspective from all of the panelists.

Who's Gonna "Win" in Wireless?

To set the stage, the looming question that I think is out there is "Can any wireless solution successfully unseat Wi-Fi® as the wireless choice for delivering high-quality video streams in either one room or multiple rooms of the house?" A lot of folks look at the hype and the eventual downfall of UWB solutions as a cautionary tale to those who want to work outside the boundaries of Wi-Fi to deliver high-performance. They argue that Wi-Fi's performance-to-price ratio is more than adequate for most connected CE use cases.

Wi-Fi's clear success to date has been on the enterprise and consumer router side of the business, but its CE business is no slouch. Both stationary (gaming consoles, digital televisions, set-top boxes, and printers) and mobile/portable (handheld games, cameras, and portable music players) consumer electronics rather than PCs provided the fastest growth for Wi-Fi chipsets in 2008, according to the Wi-Fi Alliance. Among the more than 4,000 Wi-Fi certified products, consumer electronics products – and those particularly aimed at entertainment – represent a very low percentage of the total (less than ten percent), but are a growing percentage of annual shipments, according to the Alliance.

But is Wi-Fi (and specifically 802.11n) sufficient for the needs of consumer electronics manufacturers and service providers that are seeking to facilitate the movement of several high-definition streams of content around the home? Upstarts like AMIMON and SiBEAM developed their own solutions to address this issue, and they've had some success. Noam Geri at AMIMON reminds me that they sold more than 100,000 chipsets in 2008, and their customers include:

  • Hisense wireless HDTVs.
  • SONY shipped TVs in Europe and in Japan embedded with AMIMON’s chipset, including the world’s thinnest LCD TV – The SONY ZX1 which does not have any connectors on the thin panel – only wireless.
  • In the U.S., SONY is shipping the SONY BRAVIA Wireless link based on AMIMON’s technology. This is a wireless video accessory for BRAVIA TVs (you can get it on Amazon.com).
  • Sharp is selling a wireless kit based on AMIMON chipset in Japan, and Mitsubishi is selling a wireless TV with embedded AMIMON chipset.
  • In the U.S., Gefen is selling dongles based on AMIMON’s chipset, and most recently Philips announced a wireless dongle based on AMIMON.
  • There are also customers in the professional and Medical spec like Stryker and IDX (professional wireless HD camera system).

On the SiBEAM side, products include:

  • Panasonic Z1 Series - NeoPDP HDTV;
  • Toshiba REGZA LINK;
  • LG LHX and LH85 LCD TVs; and
  • GefenTV Wireless for HDMI 60Ghz Extender

At the same time, companies like Celeno and ProVision have their own design wins. Celeno's announced customers include a Cavium Networks reference design; Comtrend Wireless Video Adapters for distribution of multiple HD IPTV content from any Ethernet-equipped home gateway or DVR to IP set top boxes; and SerComm Corp. IP902CL, a Wireless HD IPTV networking solution.

ProVision's announced customers include the AXAR Media AXAR 1000, a Sender and Receiver kit that enables HD video, from set-top boxes, DVRs or Blu-ray players, to be securely distributed over Wi-Fi.

The Compressed/Uncompressed Debate

During the panel, SiBEAM's Sheung Li was ... how shall we put this ... forceful ... in his argument that a 60 GHz uncompressed solution was the only way to provide consumer electronics manufacturers and content owners with assurances that the quality of their 1080p high-def video wouldn't be compromised. We had just started digging into this subject during the CES panel, so I wanted to better understand the sentiments of the other companies in this regard.

Lior Weiss, Celeno: Indicates that H.264 encoding is sufficient for what he calls "perceived lossless compression," where the human mind cannot distinguish the video quality of an uncompressed versus a compressed stream. He notes that encoding companies and video experts have indicated that a solution that facilitiates the distribution of 720p60 quality video at 12 Mbps or 1080p60 at 30Mbps can be considered "lossless" according to the companies with whom he's speaking. He gives credit to the 60 GHz solution providers for developing a true uncompressed solution, but questions the robustness of the 60 GHz link to provide a consistent high-quality experience. Weiss notes that an examination of the quality of the radio link is a critical factor in evaluating a solution's effectiveness, noting that compression that is otherwise not detectable by the human eye allows for a solution that can operate at a lower bitrate and enjoy higher robustness overall.

Noam Geri, AMIMON: Noam appears to agree with Lior in the sense that the design of the radio link itself is a critical factor in determining the type of quality of experience a wireless networking solution can deliver.

"Trying to deliver all the information in a brute-force approach will actually result in very poor quality because when the channel degrades the link will break and picture will be lost, and even occasional breaks in the link are very annoying," he notes. "A better approach is to design a solution that is tolerant to errors and loss due to degradation in channel conditions."

Cost Considerations

I asked our panelists to provide some perspective on the costs for implementing their solutions, and received some interesting responses.

Lior Weiss, Celeno: The Celeno solution is based on a Transmitter chipset (i.e. Access Point) and a Receiver chipset (i.e. client). The uniqueness of the Celeno technology (vs. other Wi-Fi solutions) that our powerful “sauce” can yield HD performance even if the receiver chipset is not Celeno’s. That helps us with the price curve of the solution.

Weiss notes that the Celeno solution is receiver-agnostics, so as long as the transmitter used their solution, then connected CE devices could use other Wi-Fi chipsets. The transmitter side could be an over-the-top STB bundled with the TV, it could be a Tuner in case of a “two-piece” TV model, etc.

"On the transmitter side, Weiss notes,"the Celeno chip has about 25% premium over plain-vanilla Wi-Fi. “By working on the client side with Wi-Fi price leaders we achieve end to end attractive price points, especially when multiple clients are involved." Weiss notes, however, that this wouldn't preclude Celeno from working with other Wi-Fi chipset vendors. "In fact, he notes, "keeping open architecture is a great advantage and is perceived well with the tier 1 OEM’s."

For example, a big TV manufacturer who is bundling USB Wi-Fi adaptors on the TV side with Media server device for DVR functionality complained about the performance and the reach of the solution between rooms for video streaming. After plunking in a Celeno CL1300 into the Media server side (and still with the same USB Wi-Fi adaptor) performance was improved dramatically and became a true multi-room solution. Lastly, The Celeno solution is standard Wi-Fi based. We are interoperating with other 802.11n chipsets. I would call it a 'novel' implementation of Wi-Fi."

Noam Geri, AMIMON: WHDI uses 5GHz, OFDM, MIMO – basically 80% of the WHDI solution is similar to 802.11n so the inherent costs of WHDI is about the same as 802.11n. Current offering of WHDI is more expensive than wi-fi, because of the volumes and the many cost reduction iterations that wi-fi underwent. Overtime WHDI should converge to the same price-points of wi-fi since it has similar complexity and uses the same RF technology.

It is worth noting that WHDI and Wi-Fi have different functionality so this is not an apples-to-apples comparison. WHDI can deliver uncompressed video and wi-fi chipset can deliver IP data (or compressed video). In order to build a solution based on wi-fi to deliver raw video, one would need to add a real time compression engine (e.g H.264 encoder). The compression engine not only degrades quality and adds latency, but also adds complexity and costs, so a solution to deliver uncompressed video based on wi-fi is actually be more expensive than WHDI.

In fact, the synergies between 802.11 and WHDI will lead within 2-3 years to single chip solutions that implement both WHDI and 802.11. Many systems need both. For example: a TV will need WHDI to connect to all the video sources in the home and 802.11 to connect to the Internet. With a single chip device the TV will get both for the price of one.

Power Considerations

A good deal of the analysis that focus on the wireless space target price and power as two major constraints to growth. I wanted to ask our panelists their take on the power consumption/output issues, with a specific eye on any potential regulatory concerns or the use of their solutions on more power-sensitive consumer electronics devices, such as mobile phones or portable multimedia players.

Lior Weiss, Celeno: We are fully compliant to ETSI and FCC regulations for 5GHZ radio transmission as far as power output, DFS, etc. No issue there. We are using standard power amplifier , the same as any other WiFi chipsets and transmit as per the regulated output power in the specific radio channel. Our OEM and ODM partners certify their Celeno based devices per ETSI and FCC regulations.

Noam Geri, AMIMON: WHDI is very suitable for portable applications. Current implementations of WHDI can be as lower than 2W which makes WHDI suitable for notebook computers. Granted, this is still too high for ultra-portable devices, but we can expect future versions of WHDI to enable modes of less than 1W enabling applications such as wireless portable gaming consoles and even WHDI in cell phones. This is one of the major advantages of using the 5GHz band as opposed to 60GHz which requires much higher output power levels because of the poor propagation qualities of the wireless channel.

The Fit for Wi-Fi

I received some input from Vijay Desai, the founder and CEO of ACEurity, regarding how Wi-Fi may be best positioned to support very high-quality video distribution for the home. Desai says that the use of so-called "4x4" MIMO configurations will increase coverage range and throughput. 4x4 MIMO configurations rely on four transmitting antennas and four receiving antennas that allow for four "spacial streams." Desai says that a focus on low latency will place emphasis on such enhancements as dynamic bandwidth adaption, recovery mechanisms for packet loss, and H.264 compression.

ACEurity's solution, Desai notes, is a HDQ™ (HD-ViMAC/XCoder), capable of handling real-time video at 4.5 Gbps and adapting bandwidth between 10 and 130 Mbps, even with 3 x 3 Spatial Multiplex MIMO configuration. This solution, he notes, is very resilient, capable of of withstanding greater than 50% packet loss in transit and rendering "visually losless" video with an extremly low latency (approximately one millisecond).

Desai also provided some helpful information in terms of the bill of material cost expectations for consumer electronics vendors considering a high-quality wireless video solution. He says that the desired pricing for a two-pair dongle would be $149, but with a clear path to $99. This means that the cost for the wireless solution itself should be no more than $20-$25, which he believes is achievable with volume production in 2010.

No comments: