From slow to 60GHz
Faster wireless networking is on the horizon, if we can agree how to use the 60GHz spectrum.
Consumers have embraced radio standards such as Bluetooth and Wi-Fi but their enthusiasm for sending data without wires has made the 2.4GHz band a congested place.
With so many competing radios in the same spectrum, companies are now looking much further afield to push data rates from the tens of megabits per second possible today to 1Gb/s or more. A lot of them are converging on what is currently a little-used piece of spectrum way up in the millimetre-wave band.
Most countries offer unlicensed spectrum at 60GHz, despite the fact that such signals are strongly absorbed by oxygen molecules. That puts a severe limit on the distance that 60GHz transmissions can travel, limiting it to close-range communications - unless you are willing to use a lot of power. As a result, the band has not been popular among the military and satellite-communications users that tend to occupy other parts of the millimetre-wave spectrum.
In the home, however, oxygen absorption is not a big problem, so 60GHz is beginning to look like a useful way to provide dramatic speed-ups over traditional Wi-Fi wireless local area networks (WLANs). Vendors are looking at systems that have a range of just 10m, for use within one room. Experiments with various materials have shown that it is possible to transmit signals through three walls at gigabit speeds, and for distances of up to 50m.
Different materials affect 60GHz signals in different ways, as University of Texas at Austin researchers Robert Daniels and Robert Heath found out when they tested a variety of standard building products. Plasterboard or drywall absorption of 60GHz signals is not much greater than that of 2.4GHz signals, and the loss through glass is much lower. But other materials cause a big fall-off in signal strength. This makes predicting the real-world performance of 60GHz links difficult.
The technology does have some more positive characteristics. One is that 60GHz signals bounce quite well. This could prove vital for many of the ways that people are considering using the spectrum.
Skin absorbs 60GHz radiation pretty efficiently. So walking between a computer and a media player synchronised over a 60GHz system would break the link. But with directional antennas or beam-forming techniques, devices could use walls and other objects to bounce a signal around a room, in much the same way that you can change channel on a TV by aiming the remote at the opposite wall. The other useful property is that the 60GHz spectrum is, for now, largely un-congested, unlike 2.4GHz and even 5GHz networking.
The need to use line-of-sight communication at 60GHz could also prove more of a benefit than a limitation. Although you need to find ways to steer the beam, its directionality could enable multiple nodes to use the same frequency without interfering, so long as each node is outside the cone radiated by the other transmitters.
And there is plenty of bandwidth available, even if you take the minimum subset that would enable you to use 60GHz equipment almost anywhere in the world. Australia has the smallest allocation for the 60GHz band, but it is still 3.5GHz wide. That provides plenty of space and should allow for very high-speed communications, breaking the 1Gb/s limit for wireless networking.
That's the good news. But this would not be the world of wireless communications if there weren't multiple standards vying for the same spectrum. The 60GHz band is no exception.
Not only does standards body Ecma International have the TC48 study group putting together a 60GHz standard, the IEEE 802 networking-standards organisation is close to having two competing groups. One is a point-to-point link for personal area networks (PANs) - think of it as enhanced Bluetooth - while the other is a fast lane for IEEE 802.11 wireless LANs.
The IEEE 802.15.3c task group, which is working on the PAN specification, has distributed its first draft to members and is on course to have an approved standard by September 2009. It envisions two types of communication. One, aimed at high-speed delivery of video, uses orthogonal frequency division multiplexing (OFDM) to reach, in principle, data rates of more than 3Gb/s. The second mode uses a single carrier to pass data at 1-2Gb/s, for use in portable, battery-powered devices. Terminals use a 50Mb/s control channel to negotiate which kind of data channel they use to exchange data.
The WirelessHD consortium aims to promote a cable-free form of HDMI. The companies behind WirelessHD, which include Broadcom, Denon, Onkyo, Pioneer, Sharp, Sibeam and STMicroelectronics, see the wireless network as a way of simplifying home AV set-ups. The group plans to use the IEEE 802.15.3c physical layer as the basis for its upper-layer protocols. Over a distance of up to 10m, the aim is to transmit full uncompressed 1080p video and audio, which requires a bandwidth of some 3Gb/s.
Coexistence between the various proto-standards could be problematic, as members of the IEEE 802 working groups have indicated that IEEE 802.15.3c does not yet have mechanisms to handle transmissions from other protocols using the 60GHz band.
The WirelessHD group has said it will take on some of the work of making its form of 60GHz networking work with the other standards and may also borrow some of their techniques, particularly from the 60GHz extension to Wi-Fi.
John Marshall, chairman of WirelessHD, says: "In the case of the efforts from IEEE, we think this anticipated broadening of applications in the millimetre-wave band will have a positive economic effect on 60GHz technology. WirelessHD stands to benefit from these efforts in so far as it provides new technologies for WirelessHD to consider in future generations."
Ecma's working group is developing a three-part standard that is similar to IEEE 802.15.3c at the lowest level. Although the Ecma and IEEE activities are different, there is a fair amount of overlap in the membership. GEDC, Intel, Philips Electronics and Sony all feature in the TC48 working group as well as among the membership of IEEE 802.15.3c.
Both Ecma and the IEEE settled on an approach that takes an aggregate of all the 60GHz unlicensed bands and chops them into 2.16GHz channels. This makes it possible to operate devices around the world by disabling channels that do not appear in a given country's unlicensed bands.
The territories with generous allocations could use up to three of these channels. But within those channels, there are significant differences in the way that Ecma and the IEEE have approached the physical-layer protocol. For example, IEEE 802.15.3c uses the idea of a low-bitrate common channel to act as a management layer; Ecma has opted for a peer-to-peer scheme that has units negotiate with each other how they join the network (see box, 'The European option').
The late entrant to the 60GHz party is a proposal for an extension to the 802.11 Wi-Fi protocol. This has yet to turn into an IEEE task group and may not yield products until 2011, according to James Gilb, who acts as technical editor for WirelessHD and is also active in the latest Wi-Fi efforts. It means, however, that a 60GHz version of Wi-Fi may follow hot on the heels of the delayed 802.11n standard.
According to Eldad Perahia, an Intel senior systems engineer who is pushing for the creation of a task group, the IEEE 802.11 Very High Throughput (VHT) proposal has ambitious goals.
"VHT will be the only technology that can allow a corporate or home user to roam from high-throughput, dense cells to wider-area networks in a seamless manner, while maintaining full support for the installed base security, management, diagnostics and backbone infrastructure," he wrote in a proposal to the IEEE 802.11 working group.
The proposal sees Wi-Fi extend its reach across four of the unlicensed 60GHz bands. Proponents expect chipsets to support legacy Wi-Fi at 2.4GHz and 5GHz, but move to 6GHz and 60GHz to increase massively the data rates that WLANs can support. The 6GHz layer will provide something like today's 802.11 protocol, but will hand off high-speed data connections to a mesh of point-to-point links operating at 60GHz.
Perahia says the aim is to maintain backwards compatibility with existing 802.11-series standards such as 802.11i and 802.11w for security, 802.11s for mesh networking, and 802.11k and 802.11v for network management.
Although 802.11s is a relatively new standard, both the One Laptop Per Child initiative and Intel decided to put a draft version of the mesh protocol into their cheap laptop PCs for the developing world. Intel appears to be keen on the mesh idea for 60GHz networking.
Intel decided to hold a workshop at the start of October to explore options for networks that use the 60GHz band. Presentations by Intel engineers such as Lily Yang indicate that mesh networks could extend the reach of 60GHz networks, allowing units in different rooms to communicate with each other indirectly in situations where they cannot link up directly. It looks as if the industry is taking the first of many steps on a twisting path towards high-speed networking at 60GHz.