JUST HOW GOOD IS OFDM?
(With apologies, if this has been posted...)
At a meeting of the OFDM Forum in late 2000, one speaker, Geoffrey Li of Georgia Tech, who has done significant work on OFDM technology, offered a haunting commentary on just how good a wireless modulation system OFDM is. So much so that apparently the offending commentary was scrubbed from the version of Li’s presentation that was later made public, likely because it could be perceived as “too bold.” The phrase was “OFDM is not just a modulation implied by Shannon’s information theory, it is the optimum transmission”
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Wireless Pioneer's Patience Pays Off - OFDM Developer Ready for Market Breakthrough
by Steve Stroh
Most people in the broadband wireless industry who know of Wi-LAN know it as a manufacturer of license-exempt wireless local area network (LAN) and point-to-point equipment. What’s not widely known is that Wi-LAN was founded in 1992 to develop a technology called Orthogonal Frequency Division Multiplexing (OFDM), and that it was originally intended to be used with mobile phone systems.
Wi-LAN’s problem with pushing OFDM as a wireless telephony technology was that OFDM was, at that time, mostly a concept (albeit, a concept with a lot of promise). Processing OFDM signals is incredibly computationally intensive. Although OFDM could be demonstrated, using (then) very expensive, high-speed computing systems, it would take years before such computational power would be available inexpensively, and in quantity and form factor (chips, not entire computers) that would be suitable for use in wireless telephony.
The trend was very clear — Moore’s law reigned supreme. It was apparent that the "horsepower," price and form factors required to use OFDM in wireless telephony would be available — it was just a question of when.
The mobile telephony industry eventually settled on other digital standards and moved past OFDM. That decision not to use OFDM will likely come back to haunt many in the wireless industry as it struggles with accommodating data and Internet traffic on a system designed primarily for voice. In the meantime, Wi-LAN’s OFDM technology emerges as a compelling alternative to conventional wireless telephony systems because OFDM can accommodate data.
In mid-2000, when Wi-LAN and partner Philips Semiconductor announced their success at building prototype Application Specific Integrated Circuit (ASIC) chips that incorporated Wi-LAN’s Wideband OFDM (W-OFDM) technology. Wi-LAN chose its semiconductor partner very carefully, wanting to insure that a potential partner would have a vested interest in the success of W-OFDM, not just as a supplier.
One of the reasons Philips was chosen was that Philips had previous experience with a variant of OFDM that is used for digital television broadcasting in Europe. Philips was also the "patent pool" holder of OFDM patents relating to digital broadcasting. For the company to accept Wi-LAN and its W-OFDM patents into the Philips OFDM family provided some very significant credibility for Wi-LAN.
The broadband wireless industry is increasingly recognizing that Wi-LAN’s W-OFDM technology may well be the best answer to many of the technical issues of broadband wireless. After eight years of hard work, Wi-LAN became an overnight success.
BASICS OF OFDM
OFDM, at its simplest, transmits multiple, narrow signals in parallel. Doing so combines the advantages of narrowband transmission and wideband transmission. OFDM overcomes many wireless challenges better than competing modulation systems:
• Very high spectral efficiency
• Immunity to multipath interference and noise
• High bandwidth
• Reasonable range.
It can also make use of multipath effects for non-line-of sight (NLOS) communication and use low power, allowing for extensive frequency reuse.
But, again, the downside until recently was that it was difficult and expensive to make OFDM work. That hurdle appears to be past, as Wi-LAN has begun offering products based on OFDM, and entering into partnerships with companies that also see potentially lucrative opportunities in the use of the technology.
WI-LAN PRODUCTS
When Wi-LAN ultimately pursued making use of OFDM in the early ‘90s, it entered the then-infant market for license-exempt wireless LANs and wide area networks (WANs). Although true OFDM technology would take some years to develop, Wi-LAN could apply much of its technical expertise to other products. Wi-LAN’s Hopper product line of wireless bridges became well-respected in the license-exempt wireless industry and performed well with its customer base, often being installed to eliminate the exorbitant monthly fees for high-speed (T-1) data communications circuits.
The Hopper products have evolved into the Hopper Plus product line, Wi-LAN’s mainstay product until 2000. In 2000, 12 Mbps is an unremarkable speed, but in the mid-1990s, 12 Mbps was quite fast. A somewhat unique feature of the Hopper Plus, of interest to ISPs implementing large wireless networks, is that it implements a polling protocol to eliminate collisions in a wireless network. Such collisions are typically caused by units in the network that cannot "hear" each other directly. In such cases (sometimes referred to as the hidden transmitter effect), Carrier Sense Multiple Access — Collision Detection (CSMA-CD), borrowed from Ethernet, does not function well.
One Wi-LAN product with a somewhat unique capability is the Hopper Plus 22-09, which operates in the 902 to 928 MHz license-exempt industrial, scientific and medical (ISM) band with speeds of 2.25 Mbps. The product is significant because many wireless ISPs encounter situations where 2.4 GHz equipment isn’t a good answer for wireless networking. Sometimes the problem is the range of 2.4 GHz, although that’s not often an issue with increasingly sophisticated antenna systems and sensitive receivers.
More often, a problem with 2.4 GHz is that it does not penetrate foliage very well. One potential answer to the foliage problem is to use equipment that operates at 902 to 928 MHz, but since very few manufacturers make that equipment suitable for ISP use, the typical problems are that speeds are too slow — normally 100 Kbps, far below what ISPs consider an acceptable minimum speed.
I.WILL SYSTEM
In 2000, Wi-LAN unveiled its next generation of LAN/WAN products, the I.WiLL system. I.WiLL is a carrier-class wireless broadband system that fully implements the company’s W-OFDM technology to achieve speeds of 32 Mbps using the 2.4 GHz ISM band.
Unlike many competitive systems offering such speeds, I.WiLL is intended for use in point-to-multi-point network architectures to allow wireless ISPs to service both high and low bandwidth customers using the same base station equipment. Future versions of I.WiLL technology will achieve 155 Mbps, and units that will operate in other spectrums, such as the 5.8 GHz unlicensed national information infrastructure (UNII) band, are actively under development.
At the moment, sales of the I.WiLL system are stalled in the U.S. as Wi-LAN and the FCC work through its appeal of the FCC’s denial of Part 15 type acceptance on the basis that W-OFDM does not conform to the technical specifications required for spread spectrum operation in the 2.4 GHz ISM band under Part 15 regulations. The FCC action doesn’t affect Wi-LAN’s plans for I.WiLL deployment in the 5 GHz license-exempt spectrum, since different rules are in effect for that spectrum.
OFDM FORUM
Wi-LAN clearly recognized that it had a fight on its hands to get OFDM widely accepted in the wireless industry, and sought the help of some "muscular friends." Wi-LAN formed the OFDM Forum as an independent organization (Wi-LAN is a principal member whose influence in the OFDM Forum is limited to voting rights equal to that of other principal members).
The general goal of the OFDM Forum is to begin harmonizing the use of OFDM among multiple manufacturers. The forum wisely decided against being a standards body, instead working to contribute technology and expertise on OFDM to existing wireless standards bodies such as the Institute of Electrical and Electronics Engineers (IEEE), which is responsible for the well-accepted 802.11, 802.11b, and most recently, the 802.11a standards for Wireless LANs. 802.11a, which features a data rate of 50+ Mbps and operates in various 5 GHz spectrum is notable and relevant because it specifies the use of OFDM as the modulation technique for 802.11a-compliant devices.
POWERFUL COMPETITION
Cisco Systems, through the acquisition of Clarity Wireless, is developing a variant of OFDM called Vector OFDM (V-OFDM). Cisco claims that V-OFDM does not infringe on Wi-LAN’s extensive OFDM patents (Wi-LAN claims V-OFDM is, in fact, covered by its patents) and, therefore, wants to have little to do with Wi-LAN or the OFDM Forum. So much so that Cisco has formed a competing organization to the OFDM Forum called the Broadband Wireless Internet Forum (BWIF) with its own roster of partners.
Interestingly, through its other major line of wireless business (the acquisition of Aironet), Cisco, in effect, endorses Wi-LAN’s patent position regarding W-OFDM through its use of the 802.11a Wireless LAN specification, which is based on OFDM and the 802.11a committee’s recognition of Wi-LAN’s prior art and patents.
Another twist on the patent battle between Cisco and Wi-LAN is Cisco’s pending acquisition of Radiata, a manufacturer of chipsets for 802.11a wireless LANs, which Wi-LAN has sued for patent infringement. Wi-LAN claims that a thorough legal analysis of its patent leaves little room for discussion that Wi-LAN is owed patent royalties.
PARTNERSHIPS
Wi-LAN has entered into a number of partnerships and supplier relationships with companies that intend to use its technology. One of the most notable, begun in mid-1998, is Tele2 UK, which is building out a national broadband wireless Internet access system using licensed spectrum.
Representative of smaller, growing wireless ISPs in the U.S. is Air2LAN, a fast-growing Wireless ISP based in Jackson, Mississippi, and just beginning to offer wireless broadband Internet access in Houston, Texas.
Wi-LAN’s most ambitious, and potentially lucrative, partnership is with 4G Network Technologies (4GNT) of Dallas, Texas. In short, 4GNT is actively working to deploy a "fourth generation" mobile wireless network (completely skipping the much-heralded 3G wireless network).
4GNT plans to use Wi-LAN W-OFDM technology to offer mobile and fixed broadband Internet access at 25 Mbps, the "per sector" speed shared among all users in that sector. Software upgrades expected later in 2001 will increase that speed to 90 Mbps and eventually to 155 Mbps.
4GNT plans to use the license-exempt UNII spectrum. Initially, it will use Wi-LAN W-OFDM-based PC cards for laptops and personal digital assistants (PDAs). Longer term, voice services using voice over Internet Protocol (VoIP) and other technologies will simply be other applications running on the IP-based network.
That’s a powerful combination of factors:
• Rapid construction of the network
• Use of existing wireless telephony facilities (towers)
• Use of off-the-shelf, inexpensive technology
• Network optimized for high-speed data, even while
the user is mobile
• Use of license-exempt spectrum
• Flat rate pricing of about $40 per month.
Wi-LAN has taken an equity stake in 4GNT and is its exclusive equipment supplier for wireless equipment.
If its vision of a new fourth generation mobile network is, in fact, realized, it will bring its original 1992 vision full circle — high-speed, mobile, data services driven by OFDM technology. |
