`
`APOSTOLIS K. SALKINTZIS
`THE UNIVERSITY OF BRITISH COLUMBIA
`
`ABSTRACT
`The proliferation and development of cellular voice systems over the past several years has
`exposed the capabilities and the effectiveness of wireless communications and, thus, has paved
`the way for wide-area wireless data applications as well. The demand for such applications is
`currently experiencing a significant increase and, therefore, there is a strong call for advanced
`and efficient mobile data technologies. This article deals with these mobile data technologies
`and aims to exhibit their potential. It provides a thorough survey of the most important mobile
`packet data services and technologies, including MOBITEX, CDPD, ARDIS, and the emerging
`GPRS. For each technology, the article outlines its main technical characteristics, discusses its
`architectural aspects, and explains the medium access protocol, the services provided, and the
`mobile routing scheme.
`
`and have access to external data, wireless data technology
`plays a significant part because it can offer ubiquitous con-
`nectivity, that is, connectivity at any place, any time. For this
`reason, wireless data technology can be of real value to the
`business world since computer users become more productive
`when they exploit the benefits of connectivity. The explosive
`growth of local area network (LAN) installations over the
`past several years is ample evidence of the importance placed
`on connectivity by the business world.
`
`H
`
`istorically, wireless data communications was princi-
`pally the domain of large companies with special-
`ized needs; for example, large organizations that
`needed to stay in touch with their mobile sales
`force, or delivery services that needed to keep track of their
`vehicles and packages. However, this situation is steadily
`changing and wireless data communications is becoming as
`commonplace as its wired counterpart.
`The need for wireless data communications arises partially
`because of the need for mobile comput-
`ing and partially because of the need for
`specialized applications, such as comput-
`erized dispatch services and mobile fleet
`management.
`Mobile computing, which aims to
`migrate the computing world onto a
`mobile environment, is affected primari-
`ly by two components: portability and
`connectivity. Portability, i.e., the ability
`to untether computers from the conven-
`tional desktop environment, is getting
`increasingly feasible because with the
`continuous improvement in integration,
`miniaturization, and battery technology,
`the differences in performance and cost
`between desktop and portable computers
`is shrinking. Therefore, the processing
`power of desktop computing is becoming
`available to portable environments and
`this is highly desirable as far as produc-
`tivity is concerned.
`Regarding the connectivity, i.e., the
`ability to connect to external resources
`
`n FIGURE 1. Categories of wireless data networks.
`
`2
`
`IEEE Communications Surveys • http://www.comsoc.org/pubs/surveys • Third Quarter 1999, vol. 2 no. 3
`
`Telit Wireless Solutions Inc. and Telit Communications PLC Exh. 1133 p. 1
`
`
`
`n FIGURE 2. Schematic illustration of mobile data applications.
`
`Usually, portability and connectivity are at odds: the
`more portability increases, the more difficult it becomes to
`connect to external resources. However, wireless data tech-
`nology provides the means to effectively combine both capa-
`bilities and, therefore, it is an essential technology for
`mobile computing.
`Fig. 1 presents the various wireless data technologies,
`which are essentially divided into two categories according to
`their mobility characteristics. For wide-area mobility there
`are mainly two available technologies: data transmission over
`cellular networks, whether analog or digital, and data trans-
`mission over mobile data networks. As shown in Fig. 1, the
`main difference between these two technologies is the data
`transport mode. Cellular networks, being primarily voice ori-
`ented, utilize circuit switching technology1 and, therefore, are
`optimized to isochronous data traffic conditions, whereas
`mobile data networks employ packet switching technology
`and are ideal for asynchronous data traffic transmission. Cur-
`rently, due to physical layer constraints, wide-area networks
`typically feature low-speed wireless data transmission, on the
`order of 9600 b/s. However, with the emerging new protocols,
`much higher data transmission speed is supported. For exam-
`ple, GPRS will support data transmission rates up to 115
`
`1 Some digital cellular networks (such as GSM) will utilize packet data
`service in the near future.
`
`kb/s, and HSCSD is designed to offer up to 56 kb/s over con-
`ventional voice channels.
`On the other hand, local-area wireless data networks,
`which are typically employed as private systems in businesses,
`conference rooms, university campuses, and so on, provide
`wireless data service in a small geographical area and, for this
`reason, they do not experience the same rough physical layer
`constraints of their wide-area counterparts. Therefore, they
`are capable of supporting high-speed wireless data transmis-
`sion, on the order of a few Mb/s. For local-area mobility
`there are mainly two alternatives: data transmission over
`cordless systems (e.g., over CT-2 or DECT) and over wireless
`local area networks (LANs). As indicated in Fig. 1, cordless
`systems provide circuit-switching transport service, whereas
`wireless LANs provide packet-switching transport service.
`From the previous discussion it becomes apparent that
`there are two wireless packet data technologies: mobile data
`networks, which support wide-area, low-speed service, and
`wireless LANs, which support local-area, high-speed service.
`These technologies use packet switching to transport data
`rather than circuit switching, which is typically used in cellu-
`lar or cordless networks.
`The rest of this article is devoted to mobile data technolo-
`gy. The next section provides an outline of the most impor-
`tant mobile data applications and discusses several service
`providers that operate mobile data networks worldwide. The
`following four sections discuss in detail the MOBITEX,
`
`IEEE Communications Surveys • http://www.comsoc.org/pubs/surveys • Third Quarter 1999, vol. 2 no. 3
`
`3
`
`Telit Wireless Solutions Inc. and Telit Communications PLC Exh. 1133 p. 2
`
`
`
`Application area
`
`Specific applications
`
`Mobile office
`
`Financial and retail
`communications
`
`Remote control and
`monitoring
`
`•Remote office access or database access
`•File transfer
`•Administrative control
`•Two-way communications
`•Internet browsing via the World Wide Web
`
`•Transactions such as electronic cash or fund transfers
`which, generally, do not have very high communication
`requirements
`•Card authorization at points of sale in retail outlets
`
`•Traffic and transport informatics
`•Traffic light monitoring and traffic movement
`measurements
`•Route guidance systems
`•Variable message signs on the roadside to inform drivers
`of forthcoming events or problems on the road ahead
`•Train control systems
`•Vehicle fleet management
`•Gas, water, and electricity metering systems
`•Remote monitoring and controlling of vending machines
`•General telemetry systems
`
`Alarm signaling
`
`n Table 1. Wireless packet data applications.
`
`data
`and mobile
`CDPD, GPRS,
`technologies/services, respectively, addressing
`the most significant issues of each technology.
`The final section summarizes our main conclu-
`s i o n s .
`
`MOBILE DATA APPLICAT IONS
`Circuit-switching and packet-switching can
`make a great difference in terms of transmis-
`sion cost, throughput, and service quality.
`There are some applications that are best suit-
`ed to the circuit-switching model, while others
`are best suited to the packet-switching model.
`In general, packet switching is more efficient
`and consequently less costly for “bursty” appli-
`cations that transmit small quantities of data at
`every transmission. On the other hand, circuit
`switching is more efficient for large file trans-
`missions.
`From the user’s perspective, wireless packet
`data networks (which employ packet-switching)
`offer an alternative that usually guarantees
`both cheaper and improved services in a vast
`
`Operator
`
`URL
`
`Service
`
`Equipment
`
`Encoding
`
`Coverage
`(population)
`
`Roaming
`
`Belgium/Ram
`Mobile Data
`
`Finland/Telecom
`Finland
`
`Germany/
`Detemobil
`
`http://www.ram.be
`
`Ram Mobile Data
`
`http://www.tele.fi
`
`Mobitex Network
`
`http://www.t-mobil.de
`
`Modacom
`
`Handheld terminals; Mobitex
`radio modems
`8 kbit/s
`
`Vehicle-mounted
`terminals; radio
`modems
`
`Mobitex
`1.2 kbit/s
`
`Handheld terminals;
`radio modems
`
`Datatek 6000
`9.6 kbit/s
`
`Netherlands/Ram http://www.ram.nl
`Mobile Data
`
`Ram Mobile Data
`
`Handheld terminals; Mobitex
`radio modems
`8 kbit/s
`
`Sweden/Telia
`Mobile
`
`http://www.mobitex.
`telia.com
`
`Mobitex
`
`U.K./Cognito
`
`http://www.cognito.
`co.uk
`
`Cognito Mobile Data
`Solutions
`
`U.K./Paknet
`
`http://www.vodafone.
`co.uk
`
`Paknet
`
`U.K./Ram Mobile
`Data
`
`http://www.ram.co.uk
`
`Ram Mobile Data
`
`U.K./Securicor
`Datatrak
`
`http://www.securicor.
`co.uk
`
`Datatrak Network
`
`USA/Bellsouth
`
`http://www.bellsouthwd.
`com
`
`Bellsouth Wireless
`Data
`
`Canada/Rogers
`Cantel
`
`http://www.cantel.com
`
`Rogers Cantel
`Wireless Data
`
`Vehicle-mounted
`terminals; radio
`modems
`
`Cognito Messager
`terminal; Cognito
`radio modem
`
`Mobitex
`1.2 kbit/s
`
`Cognito
`encoding
`9.6 kbit/s
`
`Handheld terminals;
`radio modems
`
`Paknet encoding
`8 kbit/s
`
`Handheld terminals; Mobitex 8 kbit/s
`radio modems
`
`Vehicle-mounted
`terminal; radio
`modems
`
`Securicor
`encoding
`10 kbit/s
`
`Handheld terminals; Mobitex 8 kbit/s
`two-way pagers;
`radio modems
`
`Handheld terminals; Mobitex 8 kbit/s
`two-way pagers;
`radio modems
`
`97%
`
`95%
`
`95%
`
`98%
`
`98%
`
`87%
`
`95%
`
`89%
`
`The Nether-
`lands, U.K.
`
`No
`
`No
`
`Belgium, U.K.
`
`No
`
`No
`
`No
`
`Belgium, The
`Netherlands
`
`Not
`disclosed
`
`No
`
`93% of
`metro areas
`
`USA
`
`Main
`metro cities
`
`Canada
`
`USA/American
`Mobile
`
`http://www.ammobile.
`com
`
`ARDIS
`
`Handheld terminals;
`two-way pagers;
`radio modems
`
`RD-LAP 4.8 kbit/s
`and 9.6 kbit/s
`
`430 markets USA, Puerto
`Rico, U.S.
`Virgin Islands
`
`n Table 2. Some of the most important wireless packet data networks and operators.
`
`4
`
`IEEE Communications Surveys • http://www.comsoc.org/pubs/surveys • Third Quarter 1999, vol. 2 no. 3
`
`Telit Wireless Solutions Inc. and Telit Communications PLC Exh. 1133 p. 3
`
`
`
`range of applications. Some of these
`applications, which effectively drive
`the market for today’s wireless
`packet data networks, are listed in
`Table 1. The list divides the appli-
`cations into four categories and lists
`the potential of wireless packet data
`technology. Fig. 2 also illustrates
`some of the essential wireless pack-
`et data applications in a schematic
`f o r m .
`The primary packet data services
`currently available for mobile appli-
`cations include ARDIS, RAM
`Mobile Data, and a number of
`other services based on cellular dig-
`ital packet data (CDPD) technolo-
`gy. Several of the most important
`mobile data networks and service
`providers are summarized in Table
`2. CDPD services, not included in
`this table, are also available [9].
`
`MOBI T EX
`MOBITEX packet data technol-
`ogy is widely accepted globally and
`is considered a true de facto s t a n-
`dard. This technology was originally
`developed by Swedish Telecom,
`now called Telia Mobitel, as a pri-
`vate mobile alarm system used by
`field personnel. However, mainly
`for economic reasons, it evolved
`into a public mobile radio service.
`Continuing development has been
`made by Eritel AB under the guid-
`ance of the MOBITEX Operators
`Association (MOA) [1] and Erics-
`son Mobile Communications AB
`[2]. Commercial operation was
`introduced in Sweden in 1986 and,
`since then, a number of networks
`have been deployed in Europe, the
`United States, and Australia [3, 4].
`Only the radio frequency differs
`depending on the country: 900 MHz
`is used mainly in the U.S. and Canada, and most other coun-
`tries operate in the 450 MHz range.
`In the United States, MOBITEX technology was intro-
`duced by RAM Mobile Data, a company that was originally
`formed in 1989 as a joint business venture between world-
`wide leaders in telecommunications, including BellSouth and
`RAM Broadcasting Corporation. Today, RAM Mobile Data
`is a wholly owned subsidiary of BellSouth, with a nationwide
`system with more than 1200 base stations installed. The ser-
`vice is provided in more than 7700 cities and towns, covering
`approximately 93 percent of America’s urban business popu-
`lation, and more than 11,000 miles of interstate highway, with
`automatic seamless roaming across all service areas. Further-
`more, additional coverage is being implemented in order to
`expand the service area in the near future.
`MOBITEX networks are either installed or being deployed
`in 19 countries on five continents, including Canada, the
`U.K., France, Sweden, Finland, Norway, Belgium, the Nether-
`lands, and Australia. The MOA oversees the specifications,
`
`n FIGURE 3. MOBITEX architecture.
`
`coordinates software and hardware development, and evolves
`the technology. The specifications are published by the MOA
`without any license or fee, thus there are many terminal sup-
`pliers and equipment developers.
`MOBITEX technology offers many critical features [5]:
`• Transparent, seamless roaming, eliminating the need for
`mobile users to “register” as they move from city to city
`or for others to know the location of a subscriber to
`send him a message.
`• Store-and-forward, to ensure messages are delivered
`regardless of the user’s location or status at the time the
`message is sent.
`• Dependability, with a proven reliability factor greater
`than 99.99 percent, ensuring accurate transmission for
`every message.
`• Interoperability and more connectivity options, offering
`access to an expanding range of options in hardware,
`connectivity, and messaging destinations.
`• Capacity to support millions of subscribers.
`
`IEEE Communications Surveys • http://www.comsoc.org/pubs/surveys • Third Quarter 1999, vol. 2 no. 3
`
`5
`
`Telit Wireless Solutions Inc. and Telit Communications PLC Exh. 1133 p. 4
`
`
`
`n FIGURE 4. Typical protocol architecture within MOBITEX.
`
`• Security that makes it virtually impossible to “tap” and
`decipher wireless data.
`
`SY S T E M DE S C R I P T I O N
`
`The MOBITEX system employs a cellular layout in order
`to provide wireless communication services to a specific geo-
`graphical area. It utilizes a hierarchical structure that may
`contain up to six levels of network nodes, depending on the
`size and the area of coverage. As shown in Fig. 3, the infra-
`structure comprises three types of nodes: base stations (base),
`local switches, and regional switches. The cells served by the
`same local switch form a service area or a subnet. In each
`service area 10 to 30 frequency pairs (called channels) are
`allocated to radio service [2]. Each base station typically uti-
`lizes from one to four channels, depending on the anticipated
`cell loading. All these channels have 12.5 kHz bandwidth and
`support a data rate of 8 kb/s. The allocated RF spectrum in
`the U.S. is 935 MHz to 940 MHz for the downlink (base to
`mobile) and 896 MHz to 901 MHz for the uplink (mobile to
`b a s e ) .
`The base stations are connected to local switches via local
`telephone facilities using either X.25 or HDLC data links.
`Similarly, the local switches are connected to higher level
`nodes (regional nodes) via long distance facilities and usually
`employ the same data link protocols. At the head of the hier-
`archy lies the main exchange, which interconnects with other
`networks. Finally, another network element, the network con-
`trol center (NCC), supports network-wide management and
`supervision functions.
`A key feature of a MOBITEX network is that message
`switching occurs at the lowest possible level (this is not the
`case for some other networks), ensuring quick response times
`and reduced backbone traffic. In other words, communication
`between two mobile users inside the same cell involves only
`the cell’s base station. If the mobile users roam in different
`cells belonging to the same services area, the message turn
`around occurs at the service area’s local switch. Only mobili-
`ty, authentication, and other signaling messages need to trav-
`el upward in order to maintain proper operation.
`Furthermore, if the link between a base station and its supe-
`rior switch is lost, the base station may still operate in
`autonomous mode, where it handles only intracell communi-
`cations. This feature is supported by Ericsson’s BRS2 base
`stations (see www.ericsson.com).
`Another important feature of MOBITEX is the possibility
`to forward one packet to a number of recipients. In order to
`efficiently utilize radio resources, the originator does not
`transmit multiple copies of the same packet but, instead, only
`one packet, which includes the desired recipient list in the
`
`header. The direct address for this packet is the MOBITEX
`network (this is a special address). The first network node
`that receives the packet will split the packet into a number of
`individual packets, each addressed to an individual recipient
`included in the original address list. Subsequently, each pack-
`et is separately routed through the wireline facilities.
`
`PR O T O C O L AR C H I T E C T U R E
`
`Fig. 4 shows a layered picture of the MOBITEX inter-
`faces. MOBITEX architecture is associated only with the first
`three layers of the OSI model. However, the three protocol
`layers of MOBITEX are not clearly mapped into the corre-
`sponding OSI layers. Layers four to seven are employed and
`controlled by the applications using the network.
`The mobile terminating unit, i.e., the radio modem, inter-
`faces with a mobile or portable terminal from one side and
`with the MOBITEX infrastructure from the other side,
`through the air-interface protocol. Both these interfaces are
`standardized by MOA and their specifications are extensively
`described in [7]. The interface between the mobile/portable
`terminal and the radio modem is either physical, or logical in
`cases where both elements are implemented in a single physi-
`cal unit. When the terminal and the radio modem are physi-
`cally apart, the MOBITEX ASynchronous Communication
`(MASC) protocol is used for their interface. This protocol
`provides reliable transfer of data to/from the radio modem
`and control and status monitoring of the modem.
`Traffic at the network layer is used to:
`• Transfer information from one subscriber (or applica-
`tion) to another, such as text messages, data messages,
`status messages, and higher protocol data messages.
`• Transfer alert messages, i.e., high-priority data traffic.
`• Transfer network-layer signaling packets, such as
`login/logout requests and terminal activated/inactivated
`n o t i f i c a t i o n s .
`Every network-layer protocol unit, called MPAK (see top
`of Fig. 5), identifies the entity (e.g., an application) that origi-
`nated it. An MPAK includes a class and a type label that
`indicate its significance and its priority level inside the pack-
`et-switched wireline backbone. For example, alert messages
`have higher priority than text messages and, in case of con-
`gestion, they may maintain the required quality of service.
`Furthermore, every MPAK indicates whether it can be stored
`in the recipient’s mailbox or not. The mailbox is a temporary
`storage that can be used to buffer packets whenever they can-
`not be delivered immediately (e.g., when the recipient is in a
`tunnel). The network forwards the mailbox contents to the
`intended recipient as soon as the recipient becomes available.
`The data link layer at the radio interface, which is called
`
`6
`
`IEEE Communications Surveys • http://www.comsoc.org/pubs/surveys • Third Quarter 1999, vol. 2 no. 3
`
`Telit Wireless Solutions Inc. and Telit Communications PLC Exh. 1133 p. 5
`
`
`
`n FIGURE 5. MOBITEX frame structures.
`
`ROSI (radio OSI), takes care of the transmission toward the
`MOBITEX infrastructure. The functionality of the data link
`layer is provided by the combined action of the base station
`and the mobile station, and includes:
`• The selection of the most suitable base station, in terms
`of communication reliability, with which to communi-
`cate. (This process is called roaming and is described in
`[ 6 ] . )
`• The retransmission of data link frames that were either
`destroyed by the mobile channel impairments or collid-
`ed with neighbor transmissions.
`• The channel access procedure.
`The data link frame structure is shown in the middle of
`Fig. 5. It consists of a series of fixed-length (20 bytes) blocks,
`each having a 16-bit CRC (cyclic redundancy check) for error
`protection. The data link layer employs a selective ARQ
`(automatic repeat request) scheme at the block level to effec-
`tively recover from transmission errors. After a frame trans-
`mission, the addressee checks the received blocks for errors.
`If all the blocks are correct, it replies with a positive acknowl-
`edgment; otherwise, it requests the retransmission of only the
`corrupted blocks. A new frame can be transmitted only after
`the previous frame has been positively acknowledged. In
`other words, a stop-and-wait ARQ scheme is employed at the
`frame level.
`The physical layer frame structure is depicted in the lower
`part of Fig. 5. It starts with a frame head that is used to
`establish frame synchronization and to uniquely identify a
`base radio station. The preamble field includes a synchro-
`nization pattern that enables all the prospective receivers to
`acquire bit synchronization and to correctly decode the rest
`of the frame. The preamble contains eight pairs of alternat-
`ing 1s and 0s. If the frame is transmitted from a base station
`the pattern starts with two 1s (i.e., 1100110011001100) where-
`as, when it is transmitted by a mobile station, the pattern
`starts with two 0s (i.e., 0011001100110011). In other words,
`the physical layer can identify if a frame comes from a base
`station or from another mobile station.
`The SYNC code word that follows is used to establish
`frame synchronization. It is important to note that every
`MOBITEX network maintains its own unique SYNC code
`word. Therefore, SYNC is used as a network identification
`number at the physical layer. If a mobile receives frames
`
`from a network that uses a different SYNC from the one
`currently selected, it will not be able to acquire frame syn-
`chronization and the received frames will be discarded at
`the physical layer. This feature does not introduce any
`problems durin g the roaming proced ure, i.e ., when a
`mobile station evaluates the communications quality of
`neighbor channels, because this evaluation is based only on
`signal strength measurements and no frame decoding is
`required [6].
`The base ID and area ID fields uniquely identify a base
`radio station in a MOBITEX network. Frames originated
`from a base will carry its own base and area ID, while frames
`originated from radio terminals will carry the base and area
`ID of the destination base. These ID fields make it feasible
`for a radio terminal to accept physical layer frames only from
`one base station (the one that has been selected by the roam-
`ing entity). If, maybe due to favorable propagation condi-
`tions, a mobile station receives frames from a distant base
`station, these frames are discarded.
`Error correction coding is performed at the physical layer.
`All bytes contained in the data link blocks are put into a
`matrix. Every byte is independently encoded using a short-
`ened (12, 8) Hamming code and the parity bits that result
`from the coding are appended to each one. Thus, for every
`eight data link bits, 12 physical layer bits are transmitted in
`order to combat the mobile channel impairments. The
`employed code can correct all single errors (inside a byte)
`with a hard decision decoding.
`The modulation scheme employed in MOBITEX is typi-
`cally Gaussian minimum shift keying (GMSK) with a modula-
`tion index of 0.5. As stated before, the RF channel spacing is
`12.5 kHz and the modulation rate is 8 kb/s.
`
`CH A N N E L AC C E S S
`
`The multiple access protocol in the MOBITEX is a varia-
`tion of the well known slotted ALOHA. A mobile terminal
`(MOB) that has traffic to send is allowed to transmit only
`during specific free cycles. These cycles (i.e., repeated time
`periods) are initiated by the base station in every cell with the
`transmission of a FREE frame on the downlink channel. A
`free cycle is composed by a number of time slots, all with
`equal length. The number and the length of the time slots in
`
`IEEE Communications Surveys • http://www.comsoc.org/pubs/surveys • Third Quarter 1999, vol. 2 no. 3
`
`7
`
`Telit Wireless Solutions Inc. and Telit Communications PLC Exh. 1133 p. 6
`
`
`
`n FIGURE 6. MOBITEX channel access activity (example).
`
`a free cycle are specified by the FREE frame that initiates
`the free cycle.
`After the reception of a FREE frame, a MOB with pend-
`ing transmission traffic chooses at random a slot and sched-
`ules its transmission at the beginning of that slot. Since no
`transmission can last longer than a time slot, a collision takes
`place only if two or more MOBs schedule their transmission
`at the beginning of the same slot. A MOB that gets ready for
`transmission somewhere in the middle of a free cycle does
`not choose a random time slot; rather, it schedules its trans-
`mission at the next time slot. This is reasonable since the
`instances that some MOBs get ready for transmission are
`expected to be uniformly distributed inside a free cycle.
`If a MOB has a large data frame to send, which exceeds
`the duration of a time slot, it sends a short access request
`message instead of the data frame itself. At the end of a free
`cycle, the base station will grant access permission to every
`mobile that has successfully sent an access request and thus,
`one after the other, all MOBs will eventually transmit their
`data frames (though outside of a free cycle) before the next
`free cycle.
`
`An example of a free cycle is illustrated in Fig. 6. After
`the FREE frame, which initiates a free cycle with six total
`slots, the base station transmits a large frame to MOB 3. At
`the same time, it receives the frames transmitted from the
`mobile stations. In this example, MOB 1 transmits a status
`message at slot 2 and MOB 2 transmits an access request at
`slot 4. Both these slot numbers (2 and 4) have randomly been
`selected, provided that MOB 1 and MOB 2 were waiting for
`transmission at the beginning of the free cycle. After the end
`of the free cycle, MOB 3 transmits an acknowledgment to
`indicate that the large frame transmitted from the base sta-
`tion was correctly received. Also, the base station acknowl-
`edges the correct reception of the status message from MOB
`1 and grants channel access to MOB 2, which afterward pro-
`ceeds to the transmission of its large data frame.
`Packet collisions are treated as channel errors: a MOB
`that transmits a packet in a FREE cycle waits for a positive
`acknowledgment from the base until the beginning of the
`next FREE cycle. If this acknowledgment is not received
`before the next FREE cycle, the MOB assumes erroneous
`transmission (due to collision or channel errors) and retrans-
`mits the same packet. This procedure
`continues until an acknowledgment is
`received from the base station.
`
`C DPD
`CDPD, cellular digital packet data,
`is a mobile data technology that per-
`mits subordinate data operation on
`the spectrum assigned to the
`Advanced Mobile Phone Service
`(AMPS). It was first introduced by
`IBM as a packet-switching overlay to
`the existing analog cellular voice net-
`work and frequencies. Later, a CDPD
`system specification [8] was created
`by a consortium of cellular carriers
`including AirTouch, McCaw Cellular,
`Southwestern Bell Mobile Systems,
`NYNEX, Ameritech, GTE, Bell
`Atlantic Mobile, and Contel Cellular.
`CDPD technology is being deployed
`by a number of cellular companies in
`the U.S., including Bell Atlantic,
`Ameritech, GTE, and McCaw Cellu-
`lar. Equipment is provided by a vari-
`ety of manufacturers.
`CDPD systems are designed to
`take advantage of the idle voice chan-
`nels of an analog cellular network,
`such as AMPS. These idle channels
`are used to transmit short data mes-
`sages and establish a packet-switching
`
`n FIGURE 7. CDPD network architecture.
`
`8
`
`IEEE Communications Surveys • http://www.comsoc.org/pubs/surveys • Third Quarter 1999, vol. 2 no. 3
`
`Telit Wireless Solutions Inc. and Telit Communications PLC Exh. 1133 p. 7
`
`
`
`n FIGURE 8. Standardized interfaces within CDPD architecture.
`
`service. In order to utilize those idle
`channels, CDPD implements a hopping
`procedure among the available cellular
`frequencies. The air interface operates
`at a raw data rate of 19.2 kb/s and pro-
`vides forward error correction to combat
`the interference and fading of the cellu-
`lar channels.
`The Wireless Data Forum [9] is an
`industry association that handles the
`shaping of the CDPD technology and
`supports the growth of the commercial
`marketplace. This forum aims to help the operators, equip-
`ment providers, and billing system developers through the
`complicated issues they face. Among other things, the forum
`tries to help all of the CDPD operators to develop inter-
`operator roaming and invoicing.
`According to the Wireless Data Forum, by the end of the
`third quarter 1998, CDPD was available in 195 markets in the
`United States: 118 metropolitan statistical areas (MSAs), 41
`rural statistical areas (RSAs), and 36 international markets,
`and was available to 53 percent of the U.S. population.
`
`SY S T E M DE S C R I P T I O N
`
`The primary elements of a CDPD network are the end
`systems (ESs) and the intermediate systems (ISs), as shown
`in Fig. 7. The ESs represent the actual physical and logical
`end nodes that exchange information, while the ISs represent
`the CDPD infrastructure that stores, forwards, and routes the
`i n f o r m a t i o n .
`There are two kinds of ESs: The mobile end system (M-
`ES), which is a device used by a subscriber to access the
`CDPD network over the wireless interface, and the fixed end
`system (F-ES), which is a common host, server, or gateway
`
`that supports or provides access to data and applications. By
`definition, the location of an F-ES is fixed, whereas the loca-
`tion of an M-ES may change.
`Typically, each M-ES consists of a mobile terminal (per-
`sonal computer, personal digital assistant, or other standard
`device), and an additional device, the radio modem, that
`attaches to the mobile terminal and manages the radio links
`and protocols. These devices usually communicate over stan-
`dard serial protocols, such as the Serial Line Internet Proto-
`col (SLIP), or the Point-to-Point (PPP) protocol.
`On the other hand, there are two kinds of ISs: a “generic”
`IS, which is simply a (IP) router that has no knowledge of
`CDPD and mobility issues, and a mobile data intermediate
`system (MD-IS), which is a specialized IS that routes mes-
`sages based on its knowledge of the current location of an
`M-ES. More specifically, a MD-IS is a set of hardware com-
`ponents and software functions that provide switching,
`accounting, registration, authentication, encryption, and
`mobility management functions. The mobility management
`software allows the switching system to track M-ESs regard-
`less of their location in the network, and allows M-ESs to use
`a single network address. The CDPD mobility management
`software follows the mobile-IP model [10], established by the
`Internet Engineering Task Force (IETF).
`Besides the ESs and the ISs, there is
`also another element called the mobile
`data base station (MDBS), which is anal-
`ogous to the AMPS base station. A
`MDBS is a combination of a computer,
`power amplifiers, and a radio transceiv-
`er. It performs no networking functions
`but it is a link-layer relay; it sends and
`receives information from the M-ESs
`and relays it back to the MD-IS. It also
`controls the radio interface and manages
`the radio communications, and monitors
`the activity on the voice network (to
`ensure that data and voice do not inter-
`fere with each other). The MDBS cre-
`ates an air link comprising two RF
`channels for forward and reverse com-
`munications with multiple M-ESs.
`Fig. 8 shows the standardized inter-
`faces used across the CDPD network.
`The M-ESs are connected to the CDPD
`network through the A-interface (the air
`interface), while the F-ESs are connected
`through the E-interface. The E-interface
`is also used to interconnect with external
`networks. Finally, the I-interface is used
`in the backbone, between the various
`ISs, and at the interconnection points
`with other CDPD networks. All these
`interfaces and the associated protocols
`are extensively described in [8].
`
`n FIGURE 9. CDPD service with interoperability between service providers.
`
`IEEE Communications Surveys • http://www.comsoc.org/pubs/surveys • Third Quarter 1999, vol. 2 no. 3
`
`9
`
`Telit Wireless Solutions Inc. a