sign in sign up
<<
Home , ALOHAnet, Norman Abramson, Reservation ALOHA, Wireless network

LYT-HISTORY-Abramson-Dec 11/19/09 12:32 PM Page 21

HISTORY OF COMMUNICATIONS EDITED BY MISCHA SCHWARTZ

INTRODUCTION BY EDITOR

The following article provides a concise, clearly written dis- amazing about this scheme is its basic simplicity, yet profound cussion of the history of the well-known Aloha protocol by Prof. ability to find application in such a widespread range of commu- Norman Abramson, the key individual responsible for its inven- nication systems. tion, development, and initial deployments in a variety of appli- The protocol, in its most concise form, says transmit at will. If cations involving joint use of a given medium by potentially interference is detected, retransmit some random time later. What interfering systems. He discusses the use of this protocol and its can be simpler? Yet, despite its apparent simplicity, such a proto- various incarnations, continuing to this day, in such disparate col was not at all obvious at the time of its invention and initial applications as wired , communications, and deployment. Enjoy reading this article! mobile systems, among others. What is, in my mind, so —Mischa Schwartz

THE ALOHANET — SURFING FOR WIRELESS DATA NORMAN ABRAMSON, UNIVERSITY OF HAWAII

INTRODUCTION sion detection (CSMA/CD) in Ethernet four-year college in Hilo, Hawaii, and The development of the Aloha System and CSMA with collision avoidance five two-year community colleges on project at the University of Hawaii (CSMA/CA) in WiFi and WiMAX. the islands of Oahu, Kauai, Maui, and beginning in the late 1960s provides his- This anecdotal account of the devel- Hawaii, all within a radius of about 300 torical insight into a variety of wireless opment of the Aloha System is based km from Honolulu. In September 1968 data network applications at the start of on an invited paper [1] first published a faculty group in the Department of the 21st century. In 1971 the Aloha Sys- in a special issue of IEEE Transactions at the University tem established and operated a UHF on Information Theory edited by Jim of Hawaii began to plan for an experi- terrestrial data network (AlohaNet) Massey in March 1985 [2]. mental -linked within the state of Hawaii. In 1973 the to connect all of these locations togeth- Aloha System used a VHF transponder BACKGROUND er in order to permit sharing of the in an experimental NASA satellite In the late 1960s a number of efforts computer resources on the main cam- (ATS-1) to demonstrate an internation- were in progress to use the existing pus. Even after the decision to build a al satellite data network (PacNet) con- worldwide network to pro- radio data network, however, a number necting NASA in California and five vide remote access to computer systems of basic design decisions dealing with universities in the , Japan, for terminals and, in some more ambi- the choice of network architecture had and Australia. Also in 1973 the Aloha tious cases, to provide connections to be addressed. From the beginning it System pioneered the unconventional among large information processing was clear that the nature of the radio use of a conventional commercial Com- systems for resource sharing. The term channel provided new system design sat channel to link the AlohaNet and “resource sharing” at that time was options not available in systems using PacNet to ARPANet in the continental often taken to mean a sharing of hard- conventional point-to-point telephone United States. ware that would be considered primi- channels. As the network planning pro- AlohaNet within Hawaii and PacNet tive by today’s standards. Nevertheless, ceeded, this key point assumed greater covering the Pacific were both based on in 1968 it was becoming apparent that and greater significance. It would be the use of a random access channel the existing circuit-switched telephone gratifying to be able to report that this architecture now generally referred to network architecture was not well suit- key difference between radio channels, as an ALOHA channel. Today ALOHA ed to the rapidly emerging data net- with and multiple access channels are utilized in all major mobile working needs of the 1970s. Indeed, it capabilities, and conventional point-to- networks and in almost all two-way would have been surprising had such a point wire (or ) channels was satellite data networks. network architecture, shaped by the appreciated as we began the project. Every time you power up your requirements of voice communications Unfortunately, such an appreciation or use that phone to at the end of the 19th century [3], been developed among the members of the establish a voice, SMS, or con- compatible with the emerging require- project only with time. As is the case in nection, the very first packet transmit- ments of networks many real-world situations, our fore- ted is sent via an ALOHA random at the end of the 20th century. The orig- sight was not as clear as our hindsight. access channel. The importance of the inal goal of the Aloha System was to Even at the beginning of the project, ALOHA channel within all wireless investigate the use of radio communica- however, it was understood that the networks has increased as the utiliza- tions as an alternative to the telephone intermittent operation typical of inter- tion of such networks for SMS and system for computer communications active computer terminals was a con- Internet traffic has increased. In addi- and to “determine those situations vincing argument against the tion, the initial Aloha System research where radio communications are prefer- assignment of point-to-point channels led directly to the work on carrier sense able to conventional wire communica- in a conventional frequency-division multiple access (CSMA) at the Univer- tions” [4]. multiple access (FDMA) or time-divi- sity of California at Los Angeles At that time the University of sion multiple access (TDMA) manner. (UCLA) which provided the theoretical Hawaii was composed of a main cam- Some more efficient form of sharing a foundation for both CSMA with colli- pus in Manoa Valley near Honolulu, a common

IEEE Communications Magazine • December 2009 0163-6804/09/$25.00 © 2009 IEEE 21 LYT-HISTORY-Abramson-Dec 11/19/09 12:32 PM Page 22

HISTORY OF COMMUNICATIONS

resource appeared necessary. The clas- sical techniques of direct sequence spreading or frequency hopping seemed to be one way of shar- ing channel resources, and a spread spectrum architecture was considered for the system. But the implementation of a spread spectrum format in each user terminal at a useful data rate and at a reasonable cost seemed a formidable task in 1969, given the exist- ing state of technology.

SYSTEM DESIGN The first year of the project was spent in the study of the above questions, the specification of a number of basic sys- tem parameters (e.g., frequency band, , data rate), and the organi- zation of a laboratory for both the radio Figure 1. ALOHA TCU 1971. and digital circuit parts of the system. In addition to me, the faculty involved in the project at that time included of Hawaii IBM computer at a data rate level it was the use of a random Thomas Gaarder, Franklin Kuo, Shu of 9600 bs anywhere within range of the access/broadcast communications archi- Lin, Wesley Peterson, and Edward Wel- radio system (about 100 km for the Alo- tecture for network communications. don. The key decision to use the direct haNet). Furthermore, we began to understand form of transmitting user information The design of the original TCU was that the advantages we were seeing in in a single high-speed packet burst in a constrained by the need to provide our network could be obtained in any shared channel, now known as an extensive debugging tools in this first network with intermittent (bursty) trans- ALOHA channel, was made at a meet- random access packet interface and by mitters as long as the bursty network ing of the project participants in 1969. the technology of the times. Some traffic flowed in a random access chan- The operation of the channel was not understanding of the distance we have nel. The imminent launch of domestic well understood at that time, and the traveled since then may be obtained by communication by the United derivation of the of an considering how one key decision in the States (Westar I was launched in 1974) ALOHA channel [4] was to come only design of the original network was suggested the use of a random access several weeks after the decision to use affected by the cost of memory. It was architecture for satellite networks, and that channel. Nonetheless, the criterion generally understood in the project that we turned our attention to the use of of design simplicity which led to that a full duplex mode of operation for the ALOHA channels in satellite systems at decision should not be underestimated TCU would be desirable from the point about that time [5, 6]. as a reliable guide to designing data of view of both system protocols and networks that work. simpler hardware design. Full duplex SATELLITE NETWORKS The remaining system design deci- operation, however, required the use of As one step in this process, it was decid- sions were easier to make. In retro- two independent buffers to store the ed to obtain a satellite link to connect spect, the only one that was crucial packets flowing into and out of the the AlohaNet in Hawaii to the rapidly from the point of view of getting some TCU, and at that time the cost of mem- expanding ARPANet packet-switched sort of acceptable system operation was ory for an additional packet buffer of network on the U.S. mainland. The the choice of frequency band. That deci- 88 bytes was about $300. We decided Aloha System funding at about this sion was made on the basis of some that full duplex operation was not time was transferred to the Advanced helpful advice from Dr. David Braver- important enough to justify that expen- Research Projects Agency (ARPA), man of Hughes Aircraft Company. Alo- diture, since we were interested in a under the direction of Dr. Lawrence haNet was assigned two 100 kHz network that might eventually contain Roberts, and the connection to the bandwidth channels at 407.350 MHz hundreds of user stations. As a point of ARPANet seemed sensible for a variety and 413.475 MHz, and, after the usual comparison, I recently purchased a of reasons. unexpected software delays, the first USB memory stick with 16 Gbytes stor- When the initial contacts with possi- ALOHA random access user terminals age for about $40 — a decrease in the ble carriers were made in order to lease went into operation in June 1971. For- price per bit of storage by a factor of a 56 kb/s satellite channel for the con- matting of the ALOHA packets as well more than 1 billion in 40 years. nection from the AlohaNet to a of as the retransmission protocols used in During 1971 and 1972, as additional the ARPANET at NASA Ames the AlohaNet were implemented by a TCUs were built and a network came Research Center in California, it was special-purpose piece of equipment into existence, our understanding of the suggested that we lease 12 satellite (designed by Alan Okinaka and David distinction between point-to-point chan- voice channels under existing tariffs for Wax) called the terminal control unit nels and random access/broadcast chan- voice grade channels, interface each of (TCU) (Fig. 1). A user terminal was nels deepened. It became clear that the these channels with a operating attached to the TCU by means of a key innovation of the AlohaNet was not at 4800 bs, and multiplex the outputs to standard RS232 interface, and the user just the first use of wireless communica- obtain the data rate desired. Such a was connected to the central University tions for a data network, but at a deeper procedure had obvious economic bene-

22 IEEE Communications Magazine • December 2009 LYT-HISTORY-Abramson-Dec 11/19/09 12:32 PM Page 23

HISTORY OF COMMUNICATIONS

an IMP at the Aloha System laboratory in Hawaii to be used with the satellite channel discussed above, I took the chalk and inserted “ the Aloha System” in his list and beside it placed the date of December 17 (chosen more or less at random). After Roberts’ return, we continued our discussion but, because of the rather long agenda, we did not discuss the installation of an IMP in Hawaii, and I forgot I had inserted an installation date of December 17 for us in the ARPA schedule on his black- board. I never did get the opportunity to discuss the installation of an IMP in our laboratory with Roberts. Instead, about two weeks before the December Figure 2. ALOHA PCU 1973. 17 date, we received a phone call from the group charged with the responsibili- fits to those providing the 12 leased versity of Alaska, Tohoku University in ty of installing the IMPs asking us to satellite channels, but it also required Sendai, Japan, the University of Elec- prepare a place for the equipment. On that ARPA lease 12 satellite voice chan- tro-communications in Tokyo, and the December 17, 1972, an IMP connecting nels to send only 56 kb/s. The leasing of University of Sydney in Australia. This the ALOHA system and AlohaNet to 12 voice channels seemed wasteful when network, called PacNet, operated at the ARPANet by means of the first we pointed out that the standard COM- 9600 b/s in an ALOHA channel using satellite channel in the ARPANet was SAT technique for transmitting a single low-cost satellite Earth stations to show delivered and installed. voice channel at that time was to trans- the potential of data networks with large By 1974, with the advent of the first form the analog voice signal into a 56 numbers of small Earth stations [7]. microprocessors by , it was clear kb/s pulse code (PCM) data Financial support for the research that much of the logic of the TCU could signal at the Earth station. In other on the Aloha System while the Alo- be handled by a microprocessor, and words, the existing tariff structure would haNet and PacNet were being built, and several new terminal controllers, based have had us convert our 56 kb/s data these two networks were connected to on the INTEL 8080, were designed by signal into 12 4800 b/s signals for trans- the ARPANet in the continental Unit- Christopher Harrison and put into mission to the satellite Earth station on ed States was provided by the Informa- operation. Because the network proto- the north shore of Oahu, where these tion Processing Techniques Office of cols could now be implemented in soft- channels would have been transformed ARPA, under the direction of Larry ware, and this software could be into 12 56 kb/s signals, transmitted to Roberts. In addition to this financial simultaneously modified for all operat- the other Earth station in California as support of the Aloha System, Roberts ing units by having the central station 12 56 kb/s signals, and transformed also contributed to the success of the broadcast new protocol parameters or back into 12 4800 b/s signals for deliv- project in a way not ordinarily obtained even a completely new protocol, these ery to the network. These signals would from funding agencies. In a real sense, new controllers were called pro- then be multiplexed to form the 56 kb/s Roberts acted as another member of grammable control units (PCUs) (Fig. data stream desired. It is a pleasure to the research staff of the project, con- 2). Once the basic protocols of the report that common sense finally pre- tributing a number of major technical ALOHA channel were analyzed and vailed, and after some discussions with results (including the first derivation of demonstrated in the AlohaNet UHF Roberts, a new 56 kb/s tariff was intro- the capacity of the slotted ALOHA radio channels, it was not long before duced by COMSAT (which showed up channel and the first analysis of capture other groups began to look into the in the COMSAT annual reports for sev- effect in ALOHA channels [8]). possibility of using ALOHA random eral years as the “ARPA tariff”) allow- Sometime in 1972, I was visiting access in other media as well. One of ing the lease of digitized 56 kb/s Roberts’ office in Washington for dis- the first and most successful of these channels as data channels. cussions dealing with both technical and efforts was begun in the doctoral disser- The ARPA/Aloha System link was administrative matters in the Aloha Sys- tation of at Harvard the first commercial satellite link to tem when he was called out of his office University published as a project MAC employ a single satellite voice channel for a few minutes to handle a minor report at the Massachusetts Institute of permitting a user to transmit 56 kb/s of emergency. 1972 was a year of rapid Technology (MIT) [9]. network data. However, this channel growth for the ARPANet as the inter- was not operated in a random access face message processors (IMPs) that COMMERCIAL APPLICATIONS mode by COMSAT, but in a conven- defined the nodes of the network were ALOHA in Cable and tional point-to-point mode. The first installed in the first network locations. Networks: Bob Metcalfe was probably network to utilize random access packet While waiting for Roberts’ return, I the first person to appreciate the signif- transmission in a satellite channel was noticed on the blackboard in his office icance of the distinction between the put into operation in the Aloha System a list of the locations where ARPA was communications architecture of point- in 1973, using the ATS-1 satellite in an planning to install IMPs during the next to-point circuit-switched and packet- experimental network that included the six-month period, together with the switched data networks and the random University of Hawaii, NASA Ames installation dates. Since I planned to access/broadcast architecture of the Research Center in California, the Uni- bring up the question of installation of AlohaNet. In his doctoral dissertation

IEEE Communications Magazine • December 2009 23 LYT-HISTORY-Abramson-Dec 11/19/09 12:32 PM Page 24

HISTORY OF COMMUNICATIONS

he coined the term “packet broadcast- packets. The initial random access pack- Europe and Japan as well) in the 1970s, ing” to emphasize this distinction, et in DOCSIS is used to set up a tempo- it was not until 1982 that the U.S. FCC although the term “random access” is rary reservation channel as in General allocated frequencies for establishing typically used to describe that commu- Service (GPRS). such a system [23]. At the beginning of nications architecture today. When his ALOHA in Satellite Networks: The the 1980s there were limited options dissertation was complete, he spent sev- first use of ALOHA random access in available for wireless transmission of eral months with the Aloha System at an operational commercial satellite net- data. Two of the most important net- the University of Hawaii, working with work took place shortly after the Ether- works of that era were ARDIS and Richard Binder who had developed and net effort in the COMSAT Marisat , both operating in the 800–900 implemented the ALOHA software and system in 1976 [17]. The request channel MHz range and both providing an ini- network protocols [10]. After leaving used to allocate voice and channels tial data rate of 8 kb/s. ARDIS was a Hawaii, Metcalfe joined the Xerox Palo in Marisat required some technique joint venture by IBM and Alto Research Center, where his devel- capable of sharing a single low-speed based on an unslotted ALOHA channel opment of Ethernet (with ) (4800 bs) channel among hundreds of and initially utilized primarily by IBM demonstrated the effectiveness of a ran- possible users, and an unslotted ALOHA service technicians; Mobitex was devel- dom access architecture on a cable- channel was chosen. Much of the theo- oped by Ericsson and Swedish Telecom based medium. The first 3 Mb/s retical foundation for both the cable- based on a slotted ALOHA channel Ethernet implemented at Xerox was based and satellite-based systems was implemented in Europe and (by RAM initially called the Alto ALOHA Net- developed at about the same time by Mobile Data) in the United States [24]. work [11]. Metcalfe then worked with Kleinrock, Tobagi, and Lam at UCLA But by the end of the 1980s it was DEC, Intel, and Xerox (the DIX group) [18–20]. Once these developments took becoming clear that effective develop- to establish an open CSMA/CD stan- place, it was astonishing how rapidly the ment of wireless data networks should dard for a 10 Mb/s Ethernet and com- acceptance of ALOHA-based medium be included within the burgeoning mercialized this standard after leaving access control (MAC) protocols pro- worldwide cellular mobile phenomenon. Xerox to found 3COM Corporation. ceeded. By 1977, at least one book had The first commercial deployment of The UHF frequencies used in the been published that referred to the a first-generation ()mobile network ALOHANet were obtained as a tempo- unslotted ALOHA channel as the “clas- in the United States was AMPS in rary experimental license authorized by sical ALOHA channel” [21, p. 587]. Chicago in 1983. AMPS was based on the FCC as part of the ARPA program Today satellites in geostationary orbit an analog architecture, but the request funding our university-based research. In provide data transmission services to channel for AMPS was implemented as the early 1970s after we had built the net- several million small Earth stations an unslotted ALOHA channel for initi- work, experimented with a wide variety of around the world. In almost all cases, ating call connections [25]. Similar net- network applications (including packet whenever these earth stations are used works were deployed in that time period repeaters, networks, and internet- for two-way data, network access is pro- in both Europe and Japan. Remember working), and developed a theoretical vided by ALOHA or slotted ALOHA. that the set of applications suitable for foundation for many of these applications Many of these networks are private net- a pervasive data network in the early [12], it was not at all clear how to con- works for companies in retailing, pub- 1980s was limited. The IBM PC was vince the appropriate regulatory authori- lishing, and the financial industry or for announced in August 1981. In Decem- ties to provide the frequencies necessary governments; other satellite data net- ber 1981 I purchased one of the first for the commercial development of wire- works using ALOHA are public net- IBM PC’s sold in Hawaii including an less ALOHA technology. works providing general IBM modem board and some primitive In the late 1970s Michael Marcus, a such as HughesNet and WildBlue. Both (but open) software to allow me to con- staff member at the U.S. Federal Com- HughesNet and WildBlue use slotted nect to the Hawaiian telephone net- munications Commission, proposed the ALOHA for traffic in the multiple access work at 300 b/s. establishment of unlicensed spread spec- channel. More recently both ViaSat and The mobile networks that started trum bands in the United States [13]. Cerona have used Spread ALOHA [22] to appear in the 1990s (GSM in Europe, After a good deal of discussion, on May for more advanced satellite networks. IS-54 and IS-95 in the United States) 9, 1985 the Commission established ALOHA in Mobile Networks: It is were digital networks which initially rules for the unlicensed use of spread convenient to characterize the introduc- used ALOHA only for the request spectrum systems to operate in the tion of ALOHA channels in wireless channel — slotted ALOHA in all industrial, scientific, and medical (ISM) mobile networks by recalling the state cases.1 The use of ALOHA packets for bands at 902–928, 2400–2483.5, and of technology, and of telecommunica- 5725–5850 MHz on a noninterfering tion regulation, during each of the past basis to other authorized users of these few decades. In the early 1970s both 1 It should be noted that the spread spectrum bands [14]. Other countries followed the ALOHANet and PacNet went into IS-95 system employed what can be called a lead of the U.S. FCC in authorizing operation, providing a clear demonstra- spread spectrum ALOHA channel in that both these bands for unlicensed transmission tion of the practicality and efficiency of spread spectrum and ALOHA properties were of data. In 1997 the IEEE 802.11 stan- this type of architec- used in the request channel for IS-95. This dard for WiFi was established using ture for user generated bursty data. spread spectrum ALOHA channel should be CSMA/CA for wireless access [15] using Later in the 1970s Ethernet and the differentiated from the Spread ALOHA channel the ISM bands. The issue of frequency Comsat Marisat network provided the [1, 22] in that IS-95 still uses different spreading availability was not an issue in the design first examples of ALOHA random sequences to separate different requesting trans- of random access for cable-based data access in commercial networks. mitters in the random access channel while the networks. The cable-based DOCSIS Although the basic concepts of cellu- Spread ALOHA channel employs the same standard [15, 16] uses a slotted ALOHA lar telephone technology had been spreading sequence for all in the channel for both Internet and voice under development at (and in random access channel.

24 IEEE Communications Magazine • December 2009 LYT-HISTORY-Abramson-Dec 11/19/09 12:32 PM Page 25

HISTORY OF COMMUNICATIONS

requests and signaling applications with- REFERENCES [19] L. Kleinrock and F. A. Tobagi, “ in Radio Channels: Part I — Carri- in the mobile voice network suggested [1] N. Abramson, “Development of the Alo- er Sense Multiple-Access Modes and Their haNet,” IEEE Trans. Info. Theory, vol. IT-31, the use of ALOHA packets for user Throughput-Delay Characteristics,” IEEE no. 2, Mar. 1985, pp. 119–23. traffic as well, so by the beginning of Trans. Commun., vol. COM-23, Dec. 1975, [2] J. Massey, “Guest Editorial,” IEEE Trans. Info. pp. 1400–16. the new century user traffic was han- Theory, vol. IT-31, no. 2, Mar. 1985, pp. [20] L. Kleinrock and S. S. Lam, “Packet Switch- dled in a variety of 2G or 2. net- 117–18. ing in a Multi-Access Broadcast Channel: [3] I. Dorros, “Retrospective — 25 Years Later,” works by GPRS [23] using a slotted Performance Evaluation,” IEEE Trans. Com- IEEE Commun. Mag., June, 2009, pp. 14–20. ALOHA random access channel com- mun., vol. COM-23, Apr. 1975, pp. 410–23. [4] N. Abramson, “The ALOHA System — Anoth- [21] J. Martin, Future Developments in Telecom- bined with a version of the Reservation er Alternative for Computer Communica- munications, 2nd ed., Prentice-Hall, 1977. ALOHA scheme first analyzed by a tions,” Proc. 1970 Fall Joint Comp. Conf., [22] N. Abramson, “VSAT Data Networks,” Proc. AFIPS Press, vol. 37, 1970, pp. 281–85. group at BBN [26]. IEEE, Special Issue on Satellite Communications, [5]. N. Abramson, “Packet Switching with Satel- The first mobile networks were vol. 78, no. 7, July, 1990, pp. 1267–74. lites,” AFIPS Conf. Proc., Nat’l. Comp. Conf., [23] M. Schwartz, Mobile Wireless Communica- introduced in Japan in late 2001 and in , June 1973, vol. 42, pp. 695–702. tions, Cambridge Univ. Press, 2005. Korea in early 2002. The first 3G [6] N. Abramson, “The Throughput of Packet [24] K. Pahlavan and A. Levesque, “Wireless Broadcasting Channels,” IEEE Trans. Commun., mobile networks in Europe began com- Data Communications,” Proc. IEEE, vol. 82, vol. Com-25, no. 1, Jan. 1977, pp. 117–28. mercial operation in 2002, while service no. 9, Sept. 1994. [7] N. Abramson, “Satellite Data Networks for [25] B. Stavenow, “Throughput-Delay Character- in the United States began in late 2003. National Development,” Telecommun. Policy, istics and Stability Considerations of the Both wideband code-division multiple vol. 8, no. 1, Mar. 1984, pp. 15–28. Access Channel in a Mobile Telephone Sys- [8] L. G. Roberts, “ALOHA Packet System With and access (W-CDMA) and pro- tem,” Proc. 1984 ACM SIGMETRICS Conf. Without Slots and Capture,” Comp. Commun. vide the higher bit rates and increased Measurement and Modeling of Comp. Sys., Rev., vol. 5, no. 2, Apr. 1975, pp. 28–42. pp. 105–12. emphasis on packet mode traffic [9] R. M. Metcalfe, “Packet Communication,” [26] W. Crowther et al., “A System for Broadcast required in the networks of today [23, MIT, Cambridge, MA, Report MAC TR-114, Communication: Reservation — ALOHA,” July 1973. 27]. ALOHA provides a natural choice Proc. 6th Hawaii Int’l. Conf. Sys. Sci., Hon- [10] R. Binder et al., “ALOHA Packet Broadcast- of network architecture for this kind of olulu, HI, Jan. 1973, pp. 371–74. ing — A Retrospect,” AFIPS Conf. Proc., [27] M. Karim and M. Sarraf, W-CDMA and traffic, and the designs of both these 1975 Nat’l. Comp. Conf., AFIPS Press, Mont- cdma2000 for 3G Mobile Networks, 3G networks reflect an increasing use vale, NJ, May 19–22, 1975. McGraw-Hill, 2002. of ALOHA random access for user [11] R. Metcalfe, Xerox PARC memo to Alto packet data as well as for signaling and Aloha Distribution on Ether Acquisition, May 22, 1973. BIOGRAPHY control purposes. [12] N. Abramson, “The ALOHA System Final NORMAN ABRAMSON [F] received an A.B. in physics As this is written on the threshold of Technical Report,” ARPA, contract no. NAS2- from Harvard College in 1953, an M.A. in 6700, Oct. 11, 1974 the 2010s, the shape of the implementa- physics from UCLA in 1955, and a Ph.D. in elec- [13] M. Marcus, “Wi-Fi and — The tion of mobile wireless is not clear, trical engineering from Stanford in 1958. He Path from Carter and Reagan-Era Faith in was an assistant professor and associate profes- but it does seem clear that an increased Deregulation to Widespread Products sor of electrical engineering at Stanford from emphasis on packet data services, Impacting Our World,” keynote talk at the 1958 to 1965. From 1967 to 1995 he was pro- Info. Economy Project Conf. on the Genesis always connected response times, and fessor of electrical engineering, professor of of Unlicensed Wireless Policy, George Mason higher data rates will lead to increased information and computer , chairman of Univ., Apr. 4, 2008; http://www.iep.gmu. the Department of Information and Computer use of ALOHA random access in 4G edu/UnlicensedWireless.php Science, and director of the ALOHA System at wireless mobile networks. [14] K. J. Negus, and A. Petrick, “History of the University of Hawaii. He is now professor Wireless Local Area Networks (WLANs) in the emeritus of electrical engineering at the Univer- Unlicensed Bands,” George Mason Univ. Law ACKNOWLEDGMENTS sity of Hawaii. He has held visiting appoint- School Conf., Info. Economy Project, Arling- ments at the University of California at Berkeley, Support for Aloha System research was ton, VA, Apr. 4, 2008. Harvard, and MIT (1965, 1966, 1980). He is a provided by the Information Processing [15] A. Tanenbaum, Computer Networks, 4th fellow of the IEC. He is also the recipient of the ed., Prentice Hall, 2003. Technology Office of ARPA within the IEEE Sixth Region Achievement Award, the PTC [16] J. Martin, “The Impact of the DOCSIS U.S. Department of Defense from 1968 20th Anniversary Award, the IEEE Koji Kobayashi 1.1/2.0 MAC Protocol on TCP,” 2nd IEEE Computers and Communications Award, the to 1974. After that date major support Conf. Consumer Commun. and Networking, IEEE Information Theory Society Golden Jubilee was provided by the U.S. National Sci- 2005, pp. 302–06. Award for Technical Innovation, the Eduard [17] D. W. Lipke et al., “MARISAT — a Maritime ence Foundation, IBM Research Labo- Rhein Foundation Technology Award, and the Satellite ,” COMSAT ratories, NEC, and Fujitsu. Additional IEEE Medal. Tech. Rev., vol. 7, no. 2, Fall, 1977. support was provided by members of [18] L. Kleinrock, Queuing Systems, Volume II: the ALOHA System Industrial Affiliates Computer Applications, Wiley, 1976, Secs. Program at the University of Hawaii. 5.11 and 5.12.

IEEE Communications Magazine • December 2009 25