`THE CASE OF WI-Fl*
`
`WOLTER LEMSTRA and Vic HAYES**
`
`Abstract
`
`In this paper we describe the genesis and development of Wi-Fi as a combined result
`of (l) a change in the US communications radio spectrum policy in the 1980s, (2) the
`industry leadership provided by NCR, its corporate successors and collaborators, to
`create a global standard and to deliver compatible p roducts under the Wi-Fi label,
`and (3) the influence of the users that moved the application of Wireless-LANs from
`the enterprise to the home, from indoor to outdoor use, from a communications
`product to a communications service, and from operators to end-users as the pro(cid:173)
`vider of that service. In concluding we assess the implications of this case for the
`formation of government policy and firm strategy. 1he case exploration and analy(cid:173)
`sis is based on contributions by experts from the field, having been involved 'first
`hand' in the innovation journey of Wi-Fi.
`
`Keywords: WLAN; IEEE 802.1 1; Wi-Fi; spectrum policy; firm strategy; sou rces of
`in novation; technology diffusion
`
`1.
`
`INTRODUCTION
`
`To-day, W i-Fi h as become the preferred means for connecting to the Internet - with (cid:173)
`out w ires: at home, in the office, in h otels, at airports, at the university campus.
`
`This paper draws upon a research project being executed within the Faculty Technology, Policy and
`Management at the Delft University of Technology (TU Delft) a imed at documenting the genesis
`and development of Wi-Fi. This is a multi-disciplinary and multi-national research project with a
`wide range of contributions from the academic community and the industry at large.
`The authors like to thank the participants of the European Communication Policy Research
`(EuroCPR) conference for the feed back on an earlier version of this paper, in particular Johan nes
`Bauer, Martin Fransm an, Anders Henten, Eli Noam. and Jean Paul Simon.
`Dr. Ir. Wolter Lemstra and Ing. Vic Hayes are Senior Research Fellow at the Section Economics of
`Infrastructures at the Faculty Tech nology, Policy and Management of the TU Delft. In their aca(cid:173)
`demic work they leverage extensive experience at the supply side of the communication industr y.
`
`Competition and Regulation in Network Industries, Volume 9 (2008), No. 2
`
`135
`
`
`
`Wolter Lemstra and Vic Hayes
`
`Increasingly Wi-Fi provides access to the Internet for remote communities in devel(cid:173)
`oping countries, e.g. in the Himalayan mountains and in the Andes. Even in rural
`areas of developed countries, for instance, in Denmark a community based Wi-Fi
`initiative emerged to provide broadband wireless Internet access, as the incumbent
`operator failed to extend the infrastructure to less profitable areas in a timely man(cid:173)
`ner.
`This is a remarkable result as wireless local area networking (WLAN) was not on
`the radar screen of the US Federal Communication Commission (FCC) when in 1980
`it initiated a market assessment that would lead to its landmark decision of 1985,
`whereby it decided to open up three radio frequency bands designated for Industrial,
`Scientific and Medical (ISM) applications for the use by radio communication sys(cid:173)
`tems, including WLANs.
`In hindsight, this should not come as a surprise. The Ethernet, which would
`become the standard for wired-LANs, was still subject of a major standardization bat(cid:173)
`tle within the IEEE in 1980. Moreover, recall that the Apple II had been launched in
`1977, while the IBM PC would be introduced in 1981, and the Internet would be named
`in 1984. Mobile computing equipment like laptops and notebooks still had to be con(cid:173)
`ceived.
`The current success of Wi-Fi is remarkable in more ways. Hitherto, the most sig(cid:173)
`nificant developments in radio frequency technology-radio-relay systems, radio and
`television broadcasting-had emerged under a licensed regime, whereby a govern(cid:173)
`ment agency provides exclusive r ights to the use of a specific part of the radio fre(cid:173)
`quency spectrum, thereby providing the application protection from interference by
`other radio frequency applications and users. The success of Wi-Fi, however, emerged
`under a license-exempt regime, whereby it had to contend with many other applica(cid:173)
`tions and users in the same radio frequency band, including micro-wave ovens and
`radar equipment.
`In this paper we will explore the innovation journey that has resulted in the global
`success of Wi-Fi, in the form of a descriptive longitudinal case study. The case starts
`in 1980 when the US Federal Communications Commission initiates a study into the
`public use of spread spectrum techniques leading to its rulemaking in 1985. We
`describe how this opportunity is us_ed by the industry, thereby focusing on the devel(cid:173)
`opments at NCR and its corporate successors to develop, market and sell a new Wire(cid:173)
`less-LAN product. The choice of NCR stems from the leading role it assumed in the
`creation and adoption of a global Wireless-LAN standard: IEEE 802.11. Subsequently
`we will explore how Wi-Fi is being deployed and shaped by the users, as part of com(cid:173)
`mercial service offerings by "hotspot" operators and through deployment as part of
`community initiatives and municipal networks. We conclude with a discussion of the
`implications of this case for government policy and firm strategy.
`
`136
`
`Intersentia
`
`
`
`Unlicensed Innovation: The Case ofWi-Fi
`
`2.
`
`TRIGGERED BY US POLICY
`
`A critical input to the development, production and application of any wireless device
`is the permission to use the radio frequency spectrum. This permission has typically
`to be granted by a government agency, as in the current spectrum management para(cid:173)
`digm the national governments have taken ownership of the frequency spectrum as a
`natural resource and assign parts of the spectrum to certain applications and users
`upon request or as a result of policy it executes (Hazlett, 2006). In the case of Wi-Fi the
`first permission is the Report and Order of May 9, 1985 of the US Federal Communi(cid:173)
`cation Comrnission1 to "[authorize] spread spectrum and other wideband emissions
`not presently provided for in the FCC Rules and Regulations" (FCC, 1985).
`The political climate was set by the Carter Administration and FCC Chairman
`Charles Ferris intended to extend the deregulation spirit to the radio frequency spec(cid:173)
`trum. He would like to end the practice whereby numerous requests for spectrum
`would be brought forward, based on special cases of technology application. The ada(cid:173)
`gio was 'let us unrestrict the restricted technologies' (Marcus, 2007; 2008). Dr. Stephen
`J. Lukasik the first Chief Scientist of the FCC, was requested to identify new commu(cid:173)
`nications technologies that were being blocked by anachronistic rules. It was Dr.
`Michael J. Marcus, employed at the Institute of Defense Analysis, who suggested that
`spread spectrum was such a technology and as a consequence was invited to join the
`FCC to follow up on the idea. In December 1979 the MITRE Corporation was invited
`to investigate the potential civil usage of spread spectrum. Their report of 1980 started
`the public consultation process on the use of spread spectrum technology.2
`
`2
`
`The Federal Communications Commission is an United States government agency, directly respon(cid:173)
`sible to Congress. The FCC was established by the Communications Act of 1934 and is charged with
`regulating interstate and international communications by radio. television. wire, satellite and
`cable. The FCC's jurisdiction covers the 50 states, the District of Columbia, and U.S. possessions
`(FCC, 2007).
`When the FCC receives petitions for new rule making, or if they see themselves a need to make a
`rules change, they have to organise a public consultation in the form of a "Notice of Inquiry, Nol".
`The public at large is invited to comment within a set period after which the public is requested to
`provide comment on comments, tl-te so-called Reply Comments.
`All comments have to be addressed in the subsequent consultation round, the so called "Notice of
`Proposed Rule Making, NPRM". In this document, the FCC also provide the proposed new rules
`with the reasons for their choices. This round is also followed by a comment and reply comment
`period.
`Again, the FCC has the obligation to address all comments and reply comments and publishes the
`results in a "Report and Order, R&O". Sometimes, a "Further Notice of Proposed Rulemaking,
`FNPRM" is included when the Order is only partially completed. A comment and reply comment
`period automatically follows the FNPRM.
`Issues found in the Order can only be appealed in Petitions for Reconsideration.
`
`Competition and Regulation in Network Industries, Volume 9 (2008), No. 2
`
`137
`
`
`
`Wolter Lemstra and Vic Hayes
`
`2.1. THE ORIGIN OF SPREAD SPECTRUM
`
`In the Notice of Inquiry the FCC proposed the civil use of spread spectrum (FCC,
`1981). Until 1981 this technique had remained officially classified as military technol(cid:173)
`ogy (Mock, 2005). The invention of spread spectrum, in the form of frequency hop(cid:173)
`ping, dates back to 1942 when a patent was granted to actress Hedy Lamarr and com(cid:173)
`poser George Antheil: U.S. Patent # 2,292,387, issued on August 11, under the title:
`"Secret Communications System". Lamarr, born as Hedwig Eva Maria Kiesler in 1913
`in Vienna, had been married to Friedrich Mandi, an Austrian arms manufacturer,
`which had exposed her to discussions on the jamming of radio-guided torpedo's
`launched from submarines. In 1937 Kiesler left Austria for America, under a contract
`with MGM. Here, she met with the composer George Antheil. Their combined insights
`in technology and music generated the idea to change the carrier frequency on a regu(cid:173)
`lar basis, akin to changing the frequency when striking another key on the piano.
`They presented their idea to the National Inventors Council and subsequently donated
`their patent to the U.S. military as a contribution to the war effort. However, the first
`practical application was after the war, in the mid 1950s, in sonobuoys used to secretly
`locate submarines (Mock, 2005 pl 1-7). The first serial production of systems based on
`direct sequence spread spectrum were most probably the Magnavox AN/ARC-SO and
`ARC-90 airborne systems. There are most probably other early systems that have
`remained classified (Marcus, 2007).
`
`2.2. THE FCC REPORT & ORDER
`
`Interestingly, the MITRE report that investigated the potential benefits, costs, and
`risks of spread spectrum communications d id not identify a strong requirement or
`need from the industry to assign spectrum for spread spectrum applications. The
`report concludes that "many potential spread spectrum applications are likely to be
`economically unattractive", other potential applications" ... may be economically fea(cid:173)
`sible, but may make poor use of the spectrum resources that they would require" and
`"[i]n certain applications, spread spectrum techniques can make more efficient use of
`the spectrum than the usual implementation of narrowband techniques ...... when
`the information bandwidth per user is low and the operating frequency is high" (Mitre
`Corp., 1980 p6-1 to 6-2). In the analysis it was recognized that spread spectrum is
`inherently more resistant to interference. The MITRE report had identified the bands
`designated for Industrial, Scientific and Medical applications (ISM bands) as bands
`" ... in which spread spectrum techniques may be able to improve the utilization of the
`spectrum ... [as these bands] are relatively unsuitable for applications requiring guar(cid:173)
`anteed high levels of performance. Indeed, since users of the ISM bands are not nom(cid:173)
`inally protected from interference, it can be argued that any productive use of these
`bands frees other spectrum resources that are needed by applications requiring pro-
`
`138
`
`I ntersentia
`
`
`
`Unlicensed Innovation: The Case ofWi-Fi
`
`tection from interference" (1980 p6-4). Typical applications in the ISM bands were
`garage door openers, retail security systems, cordless telephones and includes the
`operation of microwave ovens. Hitherto no communications applications were per(cid:173)
`mitted in the ISM bands.3
`The FCC Notice of Inquiry proposed to use spread spectrum as an "underlay"
`within other bands, i.e. sharing the frequencies with other services.4 The Notice
`triggered comments expressing fear of interference and the difficulty of tracing the
`source of interference. Based on the responses the FCC proposed two rules changes:
`one for licensed use of spread spectrum in the police bands and one for unlicensed
`use. The unlicensed proposal called for an overlay on the spectrum above 70 MHz at
`very low power (below -41 dBm) and one for unspecified power limits in the 3 bands
`designated for ISM applications (Marcus, 2007). 1he Further Notice and Notice of
`Proposed Rulemaking triggered more comments, whereby many of the respondents
`favoured the proposed authorization (FCC, 1984). Subsequently the FCC deferred all
`actions on all but the Police radio service and the use of spread spectrum in the three
`bands designated for ISM applications: the 902-926 MHz, the 2400-2483.5 MHz and
`the 5725-5850 MHz bands (FCC, 1985).5
`This FCC rulemaking that would ultimately lead to the global success of Wi-Fi had
`an interesting final twist. After the release of the spread spectrum authorization, the
`whole top leadership of the FCC Office of Science and Technology was exiled, possibly
`as a result of actions by the industry being concerned about the deregulation that
`would make the FCC less responsive to major manufacturers who wanted new tech(cid:173)
`nology only made available when it was convenient to them. An attempt was made to
`fire one deputy, and the name of the Office was changed into Office of Engineering
`and Technology. The position of Marcus was eliminated and an attempt was made to
`dismiss him from the FCC. According to Marcus: "In the months following the spread
`spectrum decision three top manager of the Office of Science and Technology were
`removed and the new organisation took no similar bold initiatives for almost a dec(cid:173)
`ade." (Marcus, 2007; 2008).
`
`3.
`
`DEVELOPED BY INDUSTRY, WITH NCR IN THE LEAD
`
`Some FCC staff members had opposed the rule changes out of fear that the new rules
`to be adopted would never be used. The reality proved otherwise. The authorizations
`
`In Europe, some communication services were permitted in the ISM bands: video surveillance by
`police, and news gathering services such as the video connections between mobile cameras on
`motorbikes and helicopters to follow the Tour de France.
`This underlay approach was similar to the approach the FCC adopted in 2003 for Ultra Wide Band
`(UWB), but in 1981 it was an idea ahead of its time (Marcus, 2007).
`The limitation on peak power was set at a level of I Watt for the three ISM bands. No limitations on
`the antenna gain were specified.
`
`Competition and Regulation in Network Industries, Volume 9 (2008), No. 2
`
`139
`
`
`
`Wolter Lemstra and Vic Hayes
`
`opened the way for innovation, because with the regulation in place companies were
`more willing to allocate investment capital to research and development. In 1988 the
`first real civil applications of spread spectrum appeared in the form of a Local Area
`Networks, e.g. Telesystems and one year later the Gambatte6 MIDI LAN, which
`became very popular with top rock musicians. A derivative of this system was used in
`nuclear power plants, under the name of Midistar - Pro. From 1990 onward the
`number of equipment authorizations by the FCC expanded significantly, see Figure 1
`(Marcus, 2000).7
`
`Figure I. Spread spectrum equipment authorizations
`
`ISM Band Spread Spectrum Annual Equipment Authorizations
`
`1000
`
`100
`
`10
`
`/
`
`--
`
`~
`
`-
`
`.~
`
`-~
`
`1988
`
`1990
`
`1992
`
`1994
`
`1996
`
`1998
`
`2000
`
`2002
`
`3.1. THE LEADING ROLE OF NCR AND ITS CORPORATE
`SUCCESSORS
`
`A leading role in the development of WLANs has been played by NCR.8 A nagging
`issue for their sales force had been the lack of mobility in the cash register product
`portfolio. Retail department stores, one of the main client groups of NCR, reconfig(cid:173)
`ured the sales floor on a regular basis and the cost of rewiring the transaction termi(cid:173)
`nals was a significant expense. To address this issue NCR had conducted a study into
`the use of infrared light technology, but quickly recognized that radio technology
`would be a much better option:" ... if it was permitted, if we could make it work, and
`
`Gambatte became Digital Wireless Corp. and then Cirronet. It was acquired by RFMonolithics in
`Texas in 2006.
`By bringing spread spectrum techniques into the civil domain, the FCC not only opened the way for
`Wi-Fi to emerge, but also facilitated the developments towards spread spectrum application in the
`field of mobile telephony in the form of CDMA, promoted by Quakom, a company established by
`Jacobs and Viterbi c.s., a month after the FCC decision (Mock, 2005).
`NCR Corporation was founded in 1879 as t he National Manufacturing Company of Dayton, Ohio,
`to manufacture and sell mechanical cash registers. In 1884 it was renamed National Cash Register
`Company. The company was acquired by AT&T in 1991. A restructuring of AT&T in 1996, led to its
`re-establishment as a separate company in 1997 (NCR, 2007).
`
`140
`
`lntersentia
`
`
`
`Unlicensed Innovation: The Case ofWi-Fi
`
`if we could turn it into affordable products" according to Don Johnson at the NCR
`Corporate R&D organisation (Johnson, 2007). The purpose and mission of Corporate
`R&D in Dayton Ohio was to (1) recognize emerging technologies and (2) to promote
`advanced development and study in areas which would benefit multiple operating
`units. All advanced development was performed in the individual operating units.
`Following the FCC Report & Order NCR Corporate initiated a feasibility study into
`the use of a wireless technology in local area networking. Copper wires, coax and
`(shielded) twisted pair, differ from radio frequency spectrum in their transmission
`properties and in the way the medium can be accessed. In terms of the Open System
`Interconnection (OSI) model this implied that new designs were required at the phys(cid:173)
`ical layer (PHY) and at the medium access layer (MAC), see also Figure 2, which shows
`the layers of the OSI protocol stack in relation to examples of current day protocols
`used in the context of the Internet (Based on Ohrtman and Roeder, 2003). Any pos(cid:173)
`sible further impact on the higher layers of the stack (network through application)
`would also have to be assessed.
`
`Figure 2. IEEE 802.11 standards mapped to the OSI reference model
`
`IEEE 802.J I related standards stack
`
`APPLICATION
`
`PRESENT A TJON
`
`SESSION
`
`TRANSPORT
`
`NETWORK
`
`IEEE802.2
`Logical Link Con1rollĀ« (I.LC)
`
`IEEE 802.11
`Media Access Control (MAC)
`
`Examples of related protocols
`used in the context of the internet
`
`HTTP, HTIPS, SMTP, POP3, IMAP, FTP,
`UUCP, NNTP. SSL, SSH, IRC, SNMP, SIP,
`RTP, Telnet, DNS
`
`TCP, UDP, SCTP, DCCP
`
`!Pv4, !Pv6, ICMP, ARP, IGMP
`
`Data Link
`Layer
`
`Ethernet, Wi-Fi, Token Ring, FDDI, PPP, ATM
`
`Infrared PHY
`
`Frequency
`Hopping
`Spreod
`Spectrum
`(FHSS)
`PHY Layer
`
`Direct
`Sequence
`Spread
`Spectrum
`(DSSS)
`PHY layer
`
`MEDlUM
`
`Physical
`Layer
`
`RS-232. EIA-422, RS-449, EIA-485, IOBaseT.
`JOOBaseT, IEEE 802. l la, IEEE 802. l lb, IEEE
`802.1 lg. DSL. ADSL
`
`Twisted pair copper, coax, fiber, radio, infra- red
`
`3.2.
`
`INVOLVEMENT OF THE DUTCH R&D CENTRE
`
`The seed money from Head Quarters in Dayton Ohio kicked off a development proc(cid:173)
`ess whereby a Dutch-based Systems Engineering centre started a feasibility study for
`
`Competition and Regulation in Network Industries, Volume 9 (2008), No. 2
`
`141
`
`
`
`Wolter Lemstra and Vic Hayes
`
`an American company to assess whether a wireless device could be developed for cash
`registers to be sold in the USA.
`The Systems Engineering centre was established to adapt the NCR products to the
`specific European requirements. The centre included a significant software develop(cid:173)
`ment team working on integration of financial systems into the IBM-world, and
`another group of experts working on adapting the telephone modem technologies to
`the European Standards. The Utr,echt Centre had become a skill centre in modem
`communication designs. One of the designs was a wired Local Area Network (MIR(cid:173)
`LAN); which NCR deployed to wire up their Cash Machines in stores before Ethernet
`became a standard.
`The choice of the Utrecht Engineering Centre for the execution of the technology
`investigation was based on their signal processing expertise, hardware design experi(cid:173)
`ence related to wired Local Area Networks, and the recent acquired radio technology
`knowledge from Philips Electronics.
`The first part of the feasibility project was to determine what power levels were
`needed and under what rules such products could be certified by the FCC. One of the
`issues was the so called "processing gain" requirements. This was the factor that had
`to be used in a spread spectrum system to "expand" the bandwidth above the band(cid:173)
`width you would "normally" need just to get your information data signal transmit(cid:173)
`ted. The logic here is that the more "spread" or processing gain the system has, the
`more the signal looks like "noise" to others - the more capable the system is in reject(cid:173)
`ing other signals, so more coexistence would be possible in a unlicensed band (Tuch,
`2007).9 Of course there is a trade off between the data rates to be achieved and the
`complexity of the total system and thus the costs. Interactions with the FCC suggested
`that a signal with a code sequence oflength 10 or greater was required. This informa(cid:173)
`tion implied that a WLAN could be realized operating at I Mbit/s or more. The team
`set to work to get the processing gain parameters set, and established a code which
`had a length of 11 with the required properties that were determined from indoor
`propagation studies.10 The feasibility study resulted in a Wireless-LAN Demo unit
`and a set of related product specifications.
`
`3.3. THE START OF PRODUCT DEVELOPMENT
`
`After the feasibility study had ended with positive results, the development team in
`Utrecht convinced the Retail Systems Division that product development was also
`best carried out by the same team. In the summer of 1987 the team set out to create a
`Wireless Network Interface Card (Wireless-NIC) to build a Wireless-LAN with an
`
`9
`
`10
`
`At the time, Bruce Tuch was leading the wireless R&D efforts of the Utrecht centre.
`The code's property: The periodic and aperiodic autocorrelation function of this 11 length code is
`"bounded" by one. Actually it turned out that this was a "known code" called the Barker Sequence
`used in Radar Systems that was "rediscovered".
`
`142
`
`lntersentia
`
`
`
`Unlicensed Innovation: The Case of Wi-Fi
`
`over-the-air data rate of 1-2 Mbit/s, to be used in the retail markets that NCR was
`serving. The NIC would have to operate in the 902-928 MHz band, the lower ISM
`band as specified by the FCC. This lower band was selected to provide the maximum
`possible range, as opposed to the ISM bands at 2.4 and 5 GHz which have higher levels
`of attenuation. Another reason was to reduce the cost of the electronics.
`The creation of a new Medium Access Control {MAC) protocol, as part of the Data
`Link Layer {DLL), was the focus of the product development effort. To limit costs and
`to reduce the development time the team intended to leverage as much as possible
`existing MAC designs and to make use of existing protocol standards where possi(cid:173)
`ble.
`
`3.4. THE ROLE OF STANDARDS WITHIN NCR
`
`Within NCR de-facto standards had been a curse rather than a blessing, as they were
`of a proprietary nature. Although the company was a leading provider of point-of-sale
`terminals, most of the time these terminals had to be connected to a back office com(cid:173)
`puting system, mostly supplied by the leading mainframe provider IBM. Having a
`dominant position in this market IBM used proprietary protocols to connect terminal
`equipment to its mainframes and mini-computers. As a result much of the protocol
`expertise of the Utrecht development team originated from the analysis and subse(cid:173)
`quent emulation of IBM protoco1s. Where NCR had the opportunity it promoted the
`use of open standards.
`
`3.5. FINDING AN EXISTING MAC PROTOCOL
`
`Finding a related MAC was in essence a search for a MAC protocol already being
`implemented using a wireless medium, or to find a MAC implemented for another
`medium, such as twisted pair copper or coax cable, that could be adapted to wireless
`use. This search led to "ALOHA'', which was one of the first Wireless Radio protocols,
`and derivates of this protocol which morphed into Ethernet and later the IEEE 802.3
`standard. While looking at the standards for LANs, another possible choice emerged:
`the Medium Access Control used in the Token Bus standard, which was very recently
`approved as IEEE 802.4. It became clear that the standards body to focus on was IEEE
`and in particular the "802" committee. The development team recognized that having
`an already established group within IEEE 802 to sponsor a new physical layer was a
`much faster process than trying to start a new standard from scratch. The IEEE 802.41
`Task Group was already working on a wireless variant driven by General Motors, but
`it seemed it was "losing steam".ll
`
`II
`
`According to the PAR this taskgroup is denoted 802.4c which through a transcription error became
`802.41.
`
`Competition and Regulation in Network Industries, Volume 9 (2008), No. 2
`
`143
`
`
`
`Wolter Lemstra and Vic Hayes
`
`The Chair of the 802.41 Task Group did not attend anymore, but the Executive
`Secretary was available and willing to convene on request of NCR a meeting in July
`1988. In the following meeting in November Vic Hayes of NCR was elected to take
`over the chair of this Task Group. However, as Tuch observed: "Making the 802.4
`protocol fit with the wireless medium was like trying to use a boat to get across a
`swamp instead of a hovercraft." (2007). Having concluded that the Token Bus MAC
`protocol was not suitable for the purpose, the MAC used as part of the IEEE 802.3
`Ethernet standard still might be adapted. One of the key issues was how to get "colli(cid:173)
`sion detect" implemented using a wireless medium. A solution developed by NCR and
`Inland Steel was presented to the IEEE 802.3 Ethernet standards group, to solicit
`interest to start a new wireless working group (Tuch and Masleid, 1991). They were
`apparently too busy on the evolution of the Ethernet standard towards higher speeds
`to support this initiative. With a negative vote for the proposal the political stage was
`set to "start from scratch" with a new Wireless MAC standard. Under the leadership
`of Bruce Tuch of NCR, the companies interested in establishing a wireless local area
`network standard quickly generated the necessary paperwork for the establishment of
`a new standardization project within IEEE. At the July Plenary meeting, the IEEE 802
`Executive Committee approved the request. With the subsequent approval by Stand(cid:173)
`ards Board the new "802.11" Working Group was born, and Vic Hayes of NCR was
`appointed as the interim chairperson.
`
`3.6. NCR TAKING THE LEAD IN IEEE 802.11
`
`September of 1990, at the first meeting of the 802.11 Working Group Vic Hayes was
`elected as the Chair.12 At the November 1991 meeting of the Work Group two Sub
`Groups were established, the MAC group and the PHY group. On a case by case basis
`the sub groups made their own rules for what materials the proponents had to submit
`for the "802.ll" membership to make a well informed decision. Once the proposals
`would be available, the two groups had the daunting task of selecting the appropriate
`technology for the project. In most of the cases the Task groups used a process of
`selection whereby in each round of voting the proposal with the lowest number of
`supporting votes would be removed from the list, until a proposal would reach major(cid:173)
`ity support. The proposal reaching majority support would be submitted to the Work(cid:173)
`ing Group for approval as the technological basis for the draft standard.
`
`12
`
`Hayes would serve as Chairperson of the IEEE 802.11 Working Group for IO years, the maximum
`period allowed.
`
`144
`
`lntersentia
`
`
`
`Unlicensed Innovation: The Case of Wi-Fi
`
`3.6.1. The first battle ground: IBM vs NCR I Symbol Technologies I Xircom
`
`The first point of contention emerging in the MAC Task Group was about the princi(cid:173)
`ple to be used in assigning capacity to a terminal based on the shared use of the radio
`spectrum. A similar issue in the Wired-LAN arena had split the industry and led to
`three different incompatible standards having been approved by the IEEE: Ethernet,
`Token Bus and Token Ring. For WLAN IBM proposed a centralized approach while
`NCR together with Symbol technologies and Xircom submitted a proposal that sup(cid:173)
`ported a decentralized mechanism. The merits of the two proposals were intensely
`debated.13 In the end the proposal for a decentralized approach won the vote; one of
`the reasons being that this protocol would support "ad hoc" networking, whereby a
`terminal would be able to independently coordinate communications with another
`terminal.
`
`3.6.2. The second battle ground: Frequency Hopping vs Direct Sequence
`
`The second area of contention was related to the PHY. In its 1985 Rule & Order the
`FCC had specified two different spread spectrum modulation techniques that could
`be used: Frequency Hopping (FHSS) and Direct Sequence (DSSS). When put to a vote
`in the PHY Task Group neither of the two modulation techniques obtained the
`required 75% level of support. Proponents of FHSS claimed it was easier to implement,
`while DSSS had the promise of a more robust system with a higher data rate. The indi(cid:173)
`viduals in the FHSS camp feared that the required investment in silicon would be
`significant, while the DSSS camp tried to refute the argument based on their experi(cid:173)
`ence in the implementation of pilot versions. As neither of the two groups could get
`the required level of support, the only way out was to include both modulation tech(cid:173)
`nologies in the standard.
`
`3.6.3. The third battle ground: HomeRF
`
`The initiative for an alternative standard called HomeRF is said to originate with
`Proxim, and led to the establishment of an industry consortium (HRFWG) in early
`1997 (Negus, Stephens et al., 2000). The main driver for this development was the
`perceived inadequate support for isochronous services, i.e. the use of telephony, in the
`IEEE 802.11 draft specification.14 The consortium adopted the Frequency Hopping
`
`13
`
`14
`
`To reach agreement within the IEEE Working Groups and Task Groups individuals opposing a
`proposal in a vote have to explain t he reasons for their opposition. By making these reasons explicit
`the group as a collective is invited to find ways to resolve the issue and if successful it has broadened
`the support for the resulting proposal.
`Companies that were involved in HomeRF development included: Butterfly Communications,
`Compaq, HP, IBM, Intel, iReady, Microsoft, Motorola, Proxim, OTC Telecom, RF Monolithics,
`Samsung and Symbionics (Lansford, 1999).
`
`Competition and Regulation in Network Industries, Volume 9 (2008), No. 2
`
`145
`
`
`
`Wolter Lemstra and Vic Hayes
`
`method as the basis for their standard.15 The HomeRF Shared Wireless Access Proto(cid:173)
`col - Cordless Access (SWAP-CA) combined portions of the OpenAir frequency hop(cid:173)
`ping PHY as developed by Proxim, CSMA/CA packet data derived from the 802.11
`Frequency Hopping standard, and TDMA-based voice support from the Digital
`Enhanced Cordless Telecommunication (DECT) standard. The FH method adopted
`by the consortium supported a data rate of 1.6 Mbit/s (Negus and Petrick, 2008).
`HomeRF was positioned as a low cost solution having a relaxed PHY specification
`supporting both isochronous (connection oriented) and asynchronous (connection(cid:173)
`less) traffic. In April 2000 Intel announced its Anypoint wireless home networking
`and in November Proxim unveiled its Symphony HRF (Palo Wireless, 2003).
`When the IEEE adopted the "802.1 lb" project for