Last week the IEEE hosted an informative panel discussion at MTTS 2010 in Anaheim titled “The Challenges, Competitions and Future Prospect of 60 GHz” that gave support to Strategy Analytics’ optimistic views on 60 GHz for short-range wireless applications. Panelists included academic and industry experts from Broadcom, IBM Research, Intel, Marvell, Panasonic, and Samsung. Aside from an interest in untethering the television, network attached storage is driving demand for faster networking, and consumers want higher bandwidth internet access and faster synchronization of ever larger multimedia files among smartphones, PCs, netbooks, portables media players and in-vehicle entertainment systems. Short-range gigabit radios can potentially serve all of these needs, and this has driven interest in commercialization and recent standardization efforts, particularly on the Wi-Fi front with the proposed 802.11ac and 802.11ad: The view among the panelists at MTTS was that WHDI (5 GHz), which has been developed explicitly for video streaming, will serve mainly in whole-house (through wall) applications. WirelessHD and WiGig will serve in cable replacement and display link applications mainly at ranges of 2 meters and less. The recently proposed 802.11ac, also known as 802.11 VHT (Very High Throughput), would use the existing 5 GHz Wi-Fi band with wide 80 MHz or 160 MHz channels, improved modulation, and simultaneous multi-user MIMO for throughputs above 1 Gbps. This would enhance Wi-Fi with faster links and extend it to new applications such as wireless connections among storage peripherals in data centers. Next would come 802.11ad, which would add a 60 GHz transport layer to Wi-Fi. According to Ali Sadri, Director of the Intel Mobile Wireless Group and Chairman of the WiGig Alliance, WiGig has been confirmed as the baseline specification for draft 802.11ad. The potential unification of Wi-Fi with 60 GHz in the form of 802.11ad would benefit both the Wi-Fi and 60 GHz communities by allowing 60 GHz to build on the installed base of Wi-Fi networks while opening new applications up to Wi-Fi. Benefits of 802.11ad, if approved, would include: • One tenth the energy per bit of 802.11ac. To transfer a 17.5 GB movie wirelessly, 802.11ac would require about 11 minutes (real data rate 780 Mbps) and use 10.5 percent of a typical cellphone battery’s energy, according to Raja Banerjea of Marvell. In comparison, 802.11ad would require less than 60 seconds and only 0.9 percent of the battery’s energy. The numbers for 802.11ad appear to work out if you assume that the battery has a capacity around 750 mA-hour at 3 volts fully charged, and that 802.11ad uses a bit less than 500 picojoules per bit at 6.7 Gbps in 60 GHz mode. • SC (single carrier) and OFDM modes for lower-rate / low power and higher-rate / high power modes respectively, combined with beam forming. Battery powered devices would use SC, which would employ pi/2 BPSK/QPSK for tolerance to phase noise and PA non-linearities, reaching 2 – 3 Gbps. In contrast, OFDM mode, using 8QPSK and above modulation schemes, would have the best multi-path tolerance and reach data rates up to 6.7 Gbps, but with higher peak-to-average ratios and higher power consumption more suited to fixed devices with wall plug power. • WiGig has developed profiles for wireless HDMI, DisplayPort, USB, and PCIe interfaces. WiGig supports HDCP 2.0 content protection. The WirelessHD Alliance has stayed a step ahead of WiGig with its latest spec, WirelessHD version 1.1, which promises maximum throughput from 10 to 28 Gbps using 60 GHz. Proponents such as chipmaker SiBeam have pointed out that WiGig as presently defined cannot support the max throughputs of HDMI 1.3 & 1.4 (10.2 Gbps) and DisplayPort 1.0 (10.8 Gbps). 60 GHz Chips. As the only commercial supplier of 60 GHz chips designed for consumer applications so far, SiBeam is benefitting from interest in WiGig as well as WirelessHD. SiBeam’s recently-announced SB8100 RF transceiver supports both WirelessHD and WiGig. Other chip makers watching 60 GHz closely and likely to offer their own chips if the market grows as expected include Atheros, Broadcom, Intel, Marvell, MediaTek (working with IBM at the moment), Samsung Electro-Mechanics, NXP, Ralink, Realtek, STMicro and TI. Qualcomm seems to have been uncharacteristically quite about 60 GHz, but the company holds or has applied for several 60 GHz-related patents including one covering extended Golay codes, which can be used for SC and OFDM coding at 60 GHz. Challenges that still stand in the way of widespread use of 60 GHz short-range radio in consumer applications include: • Common standards. It now appears that a standard encompassing Wi-Fi, WiGig and WirelessHD (and 802.15.3c by extension) will probably emerge over the next several years. WirelessHD would become the highest throughput mode of the standard. Certification of devices is still at least three years and probably more like five years away. • High power consumption, low PA efficiency. Overcoming the inherent low gain and high noise of semiconductors operating at 60 GHz requires additional power consumption compared to chips operating below 6 GHz. SiBeam’s 60 GHz chipset is said to draw up to 1 watt, worst case. Sayana Wireless LLC, a start-up in Atlanta Georgia (USA), claims to have developed a 60 GHz SiGe-BiCMOS chipset that draws only milliwatts. Unfortunately, as of May 17, 2010, founders Joy Laskar and Stephane Pinel (ECMA Vice Chair) had been suspended from Georgia Tech University following allegations of misappropriating $600,000 in university funds for use by Sayana Wireless, putting the future of the company into question (see http://www.wsbtv.com/news/23578468/detail.html.) • High-speed memory bus limitations in portable devices. Cellphones typically use memory buses running at 20 to 30 Mbps to reduce power consumption. In contrast, PCs typically have memory buses that run at or above 2 Gbps. Jeyhan Karaoguz, Technical Director at Broadcom, has stated that memory bus speed will increase to 125 Mbps or more in most portables within two years, alleviating this challenge to adoption of 60 GHz. • Global Regulations. Products that will work in Australia (59.4 – 62.9 GHz), Europe & Japan (59.0 – 66 GHz), and North America & Korea (57.0 – 64.0 GHz) using the same 60 GHz chipset will require antenna, synthesizer, LNA and PAs that can span 9 GHz, or about 15 percent bandwidth to base frequency, no small challenge compared to relatively narrow band systems operating below 6 GHz. • Testing 60 GHz chips. Testing is still a challenge. • Backhaul lagging. Backhaul links typically operate at around 100 Mbps max, imposing a bottleneck on both wired and wireless data capacity that could impede the adoption of 4G and high-speed data services, whether wired or short-range wireless. Strategy Analytics has covered short-range gigabit radio with a complete forecast in the recently published Viewpoint titled Outlook for Short-Range Gigabit Radio. The ratification of 802.11ad including WiGig would lead to a change in our forecast by effectively merging our separate forecasts for WiGig and 802.11n 4 x n and above, but it appears that our prediction of most growth in 60 GHz occurring beyond 2014 still lines up with the evidence.