`MAC Protocol Design and Performance
`
`Kin K. Leung, Bruce McNair, Leonard J. Cimini, Jr., and Jack H. Winters
`AT&T Labs - Research
`Middletown, NJ 07748
`
`Abstract — We explore the feasibility of designing an outdoor
`cellular network based on the IEEE 802.11 specification. Since
`the standard is intended for wireless local-area networks
`(WLAN), there are many technical challenges when applying the
`air interface to the outdoor environment. We study here how the
`802.11 medium access control (MAC) protocol can be applied
`and how it performs in the outdoor network. By exploiting the
`fact that timeout intervals are not explicitly specified, without
`modifying the standard, we propose a new timing structure for
`the distribution coordination function (DCF) and the handshake
`of request-to-send (RTS) and clear-to-send (CTS) to handle
`increased signal propagation delay in the outdoor network. We
`find that the DCF and RTS/CTS protocols as specified in the
`standard continue to work properly for a link distance up to 6
`km. Our analysis reveals that the DCF performance degrades
`slightly in the 802.11 network with cell size of 6 km when
`compared with the 600 m WLAN. Thus, as far as the MAC
`protocol is concerned, the 802.11 outdoor, cellular network with 6
`km cell size is feasible.
`
`INTRODUCTION
`I.
`While the wireless industry is actively developing, testing
`and deploying third generation (3G) wireless networks,
`customers are expecting services with data rate higher than
`that to be provided by 3G networks. To meet such demand for
`better services, many companies have started to provide high-
`speed data services using wireless
`local-area-networks
`(WLAN) in places such as airports and hotels. Such an
`approach is particularly attractive due to the maturity and low
`cost of the IEEE 802.11b technology [I99b, VAM99]. The
`802.11b network provides data rates up to 11 Mbps, far
`exceeding that to be offered by, for example, EDGE [SAE98,
`CQW99] and W-CDMA networks [HT00].
`
`Besides high data rates, 802.11b networks offer several
`advantages over 3G networks. First, the cost of 802.11b
`equipment is much lower than that for 3G equipment because
`of the simple design of the former networks, coupled with
`competition among WLAN vendors. Second, 802.11b
`networks operate in the 2.4 GHz ISM band, which is free
`spectrum. In contrast, the 3G spectrum is licensed and very
`expensive. Thus, both reasons make the operating cost of the
`3G network higher than that for the WLAN.
`
`On the other hand, each WLAN can serve only a small area,
`up to a few hundred meters, where a cell radius of ten
`kilometers is supported in the 3G networks. In addition, future
`3G networks are expected to provide ubiquitous coverage and
`availability. In contrast, public WLAN service is available
`only in isolated places such as airports and hotels. Users will
`
`use both types of networks, one for excellent coverage while
`the other for enhanced data rates.
`
`In this research, we explore the following question: Is it
`possible to design an outdoor, cellular network based on the
`existing 802.11 air-interface standard for wireless data
`services? If the answer is affirmative, then users can use the
`same air interface mechanism to obtain wireless services from
`indoor WLAN and outdoor 802.11 networks. There are many
`technical issues pertinent to the design of an 802.11 cellular
`network. Recall that 802.11 as well as its extension 802.11b
`[I99b] and 802.11a
`[I99a] standards were developed
`specifically for WLAN with the transmission range up to a
`few hundred meters in indoor environment. First, the signal
`propagation delay increases when applying the 802.11 to
`outdoor networks relative to the indoor WLAN, which in turn
`may affect the applicability of the medium access control
`(MAC) protocol. Second, the outdoor environment has
`increased delay spread that causes intersymbol interference.
`Further, Doppler effects due
`to mobility may require
`sophisticated processing for channel estimation.
`
`The focus of this paper is on the MAC protocol design and
`performance when using the 802.11 specification for outdoor,
`cellular networks, while radio issues will be addressed in our
`subsequent papers. Much work related to the 802.11 MAC
`protocol has been published; see e.g., [B00], [CCG00] and
`[VCM01]. The organization of the rest of this paper is as
`follows. We provide a brief description of the 802.11 MAC
`protocols in Section II. In Section III, we discuss how the
`protocols may or may not work properly in the outdoor
`networks. In addition, we estimate the maximum cell radius in
`outdoor networks due to the consideration of MAC protocols.
`Section IV analyzes the MAC performance for outdoor
`networks. Finally, our conclusion is in Section V.
`
`IEEE 802.11 MAC PROTOCOLS
`II.
`The IEEE 802.11 specification [I97] allows three kinds of
`physical layer: direct sequence spread spectrum (DSSS),
`frequency hopping spread spectrum (FHSS) and infrared (IR).
`In particular, the DSSS design supports data rates of 1 and 2
`Mbps.
`Subsequently, while maintaining
`backward
`compatibility to the DSSS 802.11 specification, the 802.11b
`was adopted to support data rates of 5.5 and 11 Mbps,
`operating in the 2.4 GHz band (the ISM band). As a result, the
`802.11b network can support 1, 2, 5.5 and 11 Mbps,
`depending on radio conditions. Another extension is 802.11a,
`which uses an entirely different physical layer known as
`orthogonal frequency division multiplexing (OFDM). 802.11a
`
`0-7803-7400-2/02/$17.00 © 2002 IEEE
`595
`
`IPR2020-00202
`Apple Inc. EX1059 Page 1
`
`
`
`illustration, station 2 does not respond, either because the
`polling message is lost or station 2 has no data to send to the
`AP. In this case, as a response is not received from station 2
`before the SIFS expires, the AP moves on to poll station 3
`within the PIFS interval, which starts from the end of the last
`polling message for station 2.
`Station 1
`
`AP
`
`AP
`
`AP
`
`can support data rates ranging from 6 to 54 Mbps, operating in
`the 5.5 GHz band (the U-NII band). It is important to note that
`it is the 802.11b networks that have been widely used recently.
`For this reason, we focus on 802.11b networks here. We also
`note that although data rates have been increased, 802.11b
`networks continue to use the original MAC protocol in the
`802.11 specification. Furthermore, the MAC protocol supports
`the independent basic service set (BSS), which has no
`connection to wired networks (i.e., an ad-hoc wireless
`network), as well as an infrastructure BSS, which includes an
`access point (AP) connecting to a wired network. The latter is
`similar to cellular networks with base stations replaced by
`AP’s. We consider only the infrastructure BSS in this paper.
`
`We provide a brief description of the 802.11 MAC protocol
`here [I97, OP99]. The 802.11 specification defines five timing
`intervals for the MAC protocol. Two of them are considered to
`be basic ones that are determined by the physical layer: the
`short interframe space (SIPS) and the slot time. The other
`three intervals are defined based on the two basic intervals: the
`priority interframe space (PIFS) and the distributed interframe
`space (DIFS), and the extended interframe space (EIFS). The
`SIFS is the shortest interval, followed by the slot time. The
`latter can be viewed as a time unit for the MAC protocol
`operations, although the 802.11 channel as a whole does not
`operate on a slotted-time basis. For 802.11b networks (i.e.,
`with a DSSS physical layer), the SIFS and slot time are 10 and
`20 µs, respectively. The slot time of 20 µs is chosen to
`account for the signal propagation and processing delays. The
`PIFS is equal to SIFS plus one slot time, while the DIFS is the
`SIFS plus two slot times. The EIFS is much longer than the
`other four intervals, and is used if a data frame is received in
`error.
`
`The 802.11 MAC supports two modes of operation: the
`Point Coordination Function (PCF) and the Distributed
`Coordination Function (DCF). The PCF provides contention-
`free access, while the DCF uses the carrier sense multiple
`access with collision avoidance (CSMA/CA) mechanism for
`contention based access. The two modes are used alternately
`in time. That is, a contention-free period by the PCF is
`followed by a contention period of the DCF.
`
`A. The PCF Protocol
`In the PCF protocol, an AP polls its associated mobile
`stations one after another by sending polling messages. If the
`AP has data to send to a mobile station being polled, the data
`can be included in the polling message. If the polled station
`has data for the AP, it is sent in the response message. When
`applicable, an acknowledgment (which acknowledges receipt
`of a previous data frame from the AP) can also be included in
`the response message.
`
`Poll +
`
`Data: S1
`
`
`Ack or
`Data
`
`Poll +
`Data: S2
`
`Poll +
`Data: S3
`
`SIFS
`PIFS
`SIFS
`(10 µs)
`(30 µs)
`(10 µs)
`
`Figure 1. The PCF of the 802.11 MAC Protocol
`
`B. The DCF Protocol
`The DCF employs the CSMA/CA mechanism and works as
`follows. A station (including the AP) with a new packet ready
`for transmission senses whether or not the channel is busy. If
`the channel is detected idle for a DIFS interval (i.e., 50 µs for
`802.11b networks), the station starts packet transmission.
`Otherwise, the station continues to monitor the channel busy
`or idle status. After finding the channel idle for a DIFS
`interval, the station: a) starts to treat channel time in units of
`slot time, b) generates a random backoff interval in units of
`slot time, and c) continues to monitor whether the channel is
`busy or idle. In the latter step, for each slot time where the
`channel remains idle, the backoff interval is decremented by
`one. When the interval value reaches zero, the station starts
`packet transmission. During this backoff period, if the channel
`is sensed busy in a slot time, the decrement of the backoff
`interval stops (i.e., is frozen) and resumes only after the
`channel is detected idle continuously for the DIFS interval and
`the following one slot time. Again, packet transmission is
`started when the backoff interval reaches zero. The backoff
`mechanism helps avoid collision since the channel has been
`detected to be busy recently. Further, to avoid channel capture,
`a station must wait for a backoff interval between two
`consecutive new packet transmissions, even if the channel is
`sensed idle in the DIFS interval. This is depicted in Figure 2.
`
`Tx starts sensing
`
`Packet
`
`Ack
`
`Packet 2
`
`DIFS
`
`SIFS
`(10 µs)
`
`DIFS + Backoff
`
`Figure 2. The DCF of the 802.11 MAC Protocol
`
`As an illustrative example in Figure 1, the AP first sends the
`polling message and data, if any, to mobile station 1 (denoted
`by S1). Station
`1
`should
`immediately
`send
`an
`acknowledgment or a data frame, if any, to the AP within the
`SIFS interval. After receiving an ACK or data from station 1,
`the AP polls mobile station 2 within the SIFS interval. In this
`
`The backoff mechanism for the DCF is an exponential one.
`For each packet transmission, the backoff time in units of slot
`time (i.e., an integer) is uniformly chosen from 0 to n-1, where
`n depends on the number of failed transmissions for the
`packet. At the first transmission attempt, n is set to a value of
`CWmin=32, the so-called minimum contention window. After
`
`596
`
`IPR2020-00202
`Apple Inc. EX1059 Page 2
`
`
`
`each unsuccessful transmission, n is doubled, up to a
`maximum value of CWmax=1024.
`
`The 802.11 specification requires a receiver to send an
`ACK
`for each packet
`that
`is successfully
`received.
`Furthermore, to simplify the protocol header, an ACK
`contains no sequence number, and is used to acknowledge
`receipt of the immediately previous packet sent. That is,
`stations exchange data based on a stop-and-go protocol. As
`shown in Figure 2, the sending station is expected to receive
`the ACK within the 10 µs SIFS interval after the packet
`transmission is completed. If the ACK does not arrive at the
`sending station within a specified ACK_timeout period, or it
`detects transmission of a different packet on the channel, the
`original transmission is considered to have failed and is
`subject to retransmission by the backoff mechanism.
`
`In addition to the physical channel sensing, the 802.11
`MAC protocol implements a network allocation vector
`(NAV), whose value indicates to each station the amount of
`time that remains before the channel will become idle. All
`packets contain a duration field and the NAV is updated
`according to the field value in each packet transmitted. The
`NAV
`is
`thus referred
`to as a virtual carrier sensing
`mechanism. The MAC uses the combined physical and virtual
`sensing to avoid collision.
`
`The protocol described above is called the two-way
`handshaking mechanism. In addition, the MAC also contains a
`four-way frame exchange protocol. Essentially, the four-way
`protocol requires that a station send to the AP a special,
`Request-to-Send (RTS) message, instead of the actual data
`packet, after gaining channel access through the contention
`process described above. In response, if the AP sees that it is
`appropriate, it sends a Clear-to-Send (CTS) message within
`the SIFS interval to instruct the requesting station to start the
`packet transmission immediately. The main purpose of the
`RTS/CTS handshake is to resolve the so-called hidden
`terminal problem.
`
`III. MAC PROTOCOLS IN OUTDOOR NETWORKS
`
`A. The PCF Protocol Infeasible
`It is important to emphasize that the SIFS and PIFS timing
`requirements for the PCF in Figure 1 are clearly defined in the
`standard. In particular, the most stringent requirement is that
`the ACK has to be received from the polled station to the AP
`within the SIFS interval, which is 10 µs for 802.11b networks.
`When the standard is used for outdoor, cellular networks, the
`distance between a mobile station and its AP is expected to be
`longer than that in the WLAN. Consider a link distance of 1.5
`km as an example. The round-trip signal propagation delay for
`the 1.5 km distance requires 10 µs. Since at least several µs
`are needed for signal processing at the receiver, the link
`distance is likely to be limited to hundreds of meters, as in
`WLAN environments. In fact, this is the intention of the
`802.11 specification. Thus, it is unrealistic to expect that the
`PCF can be supported for 802.11 outdoor networks with cell
`radius of several km.
`
`B. Applicability of the DCF Protocol
`Let us consider the DCF in the outdoor networks. It is worth
`noting that as far as the MAC protocol is concerned, the major
`difference between 802.11 outdoor networks and their WLAN
`counterparts is increased signal propagation delay. As shown
`in Figure 2, the major constraint for the applicability of the
`DCF in outdoor networks is that the ACK is expected to be
`received within the SIFS interval (10 µs) after packet
`transmission. That is, the 10 µs includes the round-trip signal
`propagation and processing at the receiver. However, in order
`to be useful, we aim at having an outdoor cell size of several
`km. Thus, the one-way signal propagation delay can be more
`than 10 µs, even neglecting the return propagation and
`processing time. Evidently, this would not be practical without
`violating the protocol specification. Our solution is based on
`the following key observation: Typically,
`there
`is no
`consequence if the ACK is received later than the SIFS
`interval. This is because, after a station transmits a packet, it
`starts an ACK_timeout period, which is not specified in the
`standard and is usually chosen to be a value much larger than
`10 µs by vendors. Thus, as long as the ACK is received before
`the timeout expires, the MAC protocol responds properly.
`
`the
`that
`implementations, we assume
`typical
`in
`As
`ACK_timeout period is longer than the DIFS interval of 50 µs.
`Then, we argue that as long as the ACK arrives at the sending
`station within
`the DIFS
`interval
`following a packet
`transmission, the DCF operates properly in the outdoor
`network environment where the link distance can reach as
`much as several km. The reasoning is as follows. First,
`because the ACK is received within the DIFS interval, the
`ACK_timeout has not expired so that the protocol can respond
`upon receipt of the ACK as if it were received within the SIFS
`interval, as originally specified in the protocol standard.
`Second, since the DCF protocol requires any station to sense
`the channel being idle for at least the DIFS interval before
`transmitting, the return of the ACK within the DIFS interval
`following the previous packet transmission by the sending
`station prevents any stations other than the receiving one from
`gaining access to the channel. Consequently, the channel is
`implicitly “reserved” for the receiving station to send the
`ACK. In addition, the pairing of a packet transmission and its
`ACK transmitted in sequence for any pair of sending and
`receiving stations remains
`intact, as required by
`the
`specification.
`
`Extending the arrival delay of ACK from the SIFS to the
`DIFS interval comes with a penalty. That is, the computation
`of the NAV assumes that the ACK returns within the SIFS
`interval. So,
`the delay extension causes an erroneous
`determination of the NAV, thus incorrect virtual sensing.
`Nevertheless, since protocol operations are based on both
`physical and virtual channel sensing, as long as the former
`works properly, the malfunctioning of the virtual sensing due
`to incorrect NAV value causes no apparent, negative impacts.
`
`Actually, the extension of the ACK arrival delay from the
`SIFS interval to the DIFS interval can also be applied to the
`RTS and CTS handshake for resolving the hidden terminal
`
`597
`
`IPR2020-00202
`Apple Inc. EX1059 Page 3
`
`
`
`transmitting. The packet transmission time is assumed to be
`constant L µs. Consider the channel activity for a successful
`packet transmission. The channel is idle for d µs and followed
`by packet transmission of L µs. As Figure 3 shows, the
`transmitter waits for d µs (DIFS interval) for the ACK. Let the
`ACK transmission time be c µs. The channel is sensed idle
`again by all stations a µs after the ACK transmission.
`
` d
`
`L
`
` d
`
`L
`
`a
`
`a
`
`Y
`
` d
`
`L
`
`a
`
`Figure 4. Busy period with Collided Transmissions
`
`Figure 4 shows a typical busy period with collided
`transmissions due to the vulnerable period for the CSMA
`protocol, where Y denotes the time span between the first and
`the last packet transmissions in the busy period. Using the
`result in [K76], the average duration of Y is given by
`−−−=
`aG
`1
`Y
`a
`.
`The average length of a busy period (which contains a
`successful transmission or collisions) is given by
`−
` aGec)
`
`++++=
`+
`YdB
`d
`aL
`(
`(2)
`where the last term accounts for the waiting and transmission
`time of the ACK for successful transmission with probability
`aGe−
`, based on the Poisson assumption of aggregated traffic.
`By the same assumption, the average cycle time, consisting of
`a busy period and the following idle period, is given by
`−
` aGecd )
`
`++++=
`+
`+
`1
`aLYdT
`(
`G
`The channel throughput S is defined as the fraction of time at
`which data is successfully transmitted. Thus, we have
`−
`eL aG
`=
`S
`
`T
`where the numerator is the average amount of time when data
`is transmitted without collision and T is obtained from (3).
`
`(1)
`
`Ge
`
`(3)
`
`(4)
`
`Three common packet sizes of 60 bytes (e.g., TCP ACK),
`576 bytes (typical size for web browsing) and 1500 bytes (the
`maximum size for Ethernet) plus a 34 byte 802.11 MAC
`header are considered. For an 802.11 network with a 1 Mbps
`data rate, the corresponding transmission time L is 0.75, 4.88
`and 12.27 msec, respectively. The sensing idle time of the
`DIFS interval of 50 µs and the transmission time c for the 112-
`bit ACK is 0.112 µs. Based on our discussions above, the link
`distance is assumed to be 6 km, and thus the one-way
`
`problem. Specifically, a sending station starts a CTS_timeout
`period after sending an RTS. The MAC protocol specifies that
`the CTS, if any, is supposed to arrive from the receiving
`station within the SIFS interval (10 µs). However, similar to
`the ACK_timeout, the CTS_timeout period is typically chosen
`to be much longer than 10 µs by equipment manufacturers.
`Therefore, by the same arguments discussed above, the arrival
`delay for the CTS can be extended to the DIFS interval.
`
`A. Maximum Cell Size for the DCF Protocol
`With the arrival delay for the ACK and CTS extended to the
`DIFS interval, let us consider its limit on the maximum cell
`size (i.e., link distance) in outdoor 802.11 networks.
`
`Tx starts sensing
`
`DIFS (50 µs)
`
`Pkt/RTS
`
`Ack/CTS
`
`DIFS (50 µs)
`
`20µs
`
`20µs
`10µs
`
`Figure 3. Allocation of ACK/CTS delay
`
`Recall that the ACK and CTS arrival delay consists of a
`round-trip signal propagation delay and signal processing
`time. As shown in Figure 3, one reasonable allocation of the
`50 µs DIFS delay is: a one-way signal propagation delay of 20
`µs and a processing time of 10 µs at the receiving station. The
`latter should not cause a processing burden for the receiver
`because the original delay of the SIFS interval is 10 µs. For
`the 20 µs propagation delay, the maximum cell size is about 6
`km. In other words, with the cell size of 6 km or less, the DCF
`protocol operates properly in 802.11 cellular networks.
`
`IV. DCF PERFORMANCE IN 802.11 OUTDOOR NETWORKS
`We present an approximate analysis of the DCF throughput
`for outdoor networks and WLAN. As shown in Figure 3, if a
`station with a packet for transmission senses the channel idle
`for the DIFS interval (denoted by d in µs in the following), it
`starts to transmit. Following the packet transmission, the
`channel remains idle for the DIFS interval and then the ACK
`is transmitted by the receiver. If the sending station senses the
`channel busy,
`it goes
`through
`the backoff mechanism
`discussed above. For simplicity, we do not model the details
`of the backoff algorithm. Instead, it is assumed that the
`aggregated
`traffic, which
`includes new packets and
`transmission reattempts, from all stations forms a Poisson
`process with an intensity of G packets/µs. This assumption is
`reasonable if the backoff period is sufficiently long so that
`new transmission and reattempts become independent sources.
`
`For simplicity, assume that the signal propagation delay a
`in µs is identical between any pair of stations. Thus, the
`vulnerable period is also given by a , during which a new
`packet transmission cannot be sensed by other stations. As a
`result, these stations under the CSMA protocol can possibly
`start their own transmissions and cause collisions. Each station
`senses the channel idle for d µs (DIFS interval) before
`
`598
`
`IPR2020-00202
`Apple Inc. EX1059 Page 4
`
`
`
`propagation delay a is 20 µs. For comparison, we also
`consider a WLAN with a service radius of 600 m with a signal
`propagation delay of 2 µs. In this WLAN, after packet
`transmission, a station waits for the SIFS interval of 10 µs as
`in the standard, instead of the DIFS interval as shown in
`Figure 3, for the arrival of the associated ACK.
`
`Applying these parameters to (1) to (4), we obtain in Figure
`5 the MAC throughput as a function of the aggregated traffic
`load for selected packet lengths. As expected, when the link
`distance increases from 600 m to 6 km for a given packet
`length, the maximum throughput decreases because of the
`increased signal propagation delay and thus the vulnerable
`period. For the 576-byte packet size, the maximum throughput
`drops from 92.9% to 84.8%, when the link distance increases
`from 600 m to 6 km. Nevertheless, since a 576-byte size is
`typical for popular web applications, the throughput of 84.8%
`is still satisfactory. For 1500-byte packets, the channel
`throughput for the 6 km cell can reach a maximum of 90.8%.
`Even for the short TCP ACKs of 60 bytes long, the channel
`throughput is about 60%. In summary, the MAC throughput is
`still satisfactory despite the increase of cell size to 6 km.
`
`Figure 5. MAC Throughput Comparison.
`
`V. CONCLUSION AND FUTURE WORK
`We have studied how the 802.11 MAC can be applied and
`how it performs in outdoor networks. By exploiting the fact
`that timeout intervals are not explicitly specified, without
`modifying the standard, we have proposed a new timing
`structure for the distribution coordination function (DCF) and
`the handshake of request-to-send (RTS) and clear-to-send
`(CTS) to handle increased signal propagation delay in the
`802.11 outdoor network. It was found that the DCF and
`RTS/CTS protocols as specified in the standard continues to
`work properly if the cell radius is less than 6 km. Our analysis
`reveals that the DCF performance degrades slightly for a cell
`size of 6 km when compared with the 600 m WLAN. Thus, as
`far as the MAC protocol is concerned, the 802.11 cellular
`network with a cell size of 6 km is feasible.
`
`In terms of future work, a major issue is to examine and
`enhance the 802.11 radio design so that it performs properly in
`the cellular environment. In a companion paper [CLMK01],
`we shall address the issue of radio link performance in the
`
`599
`
`investigate
`to
`802.11 cellular network. We also plan
`techniques such as advanced equalizers, smart antennas and
`call admission control to further improve the performance of
`the outdoor 802.11 cellular networks.
`
`ACKNOWLEDGMENTS
`We would like to thank Martin Clark, Peter Driessen, Zoran
`Kostic and Stefan Muller-Weinfurtner for their discussion.
`
`REFERENCES
`[B00] G. Bianchi, “Performance Analysis of the IEEE 802.11
`Distributed Coordination Function,” IEEE J. Selected
`Areas in Commun., Vol. 18, March 2000, pp. 535-
`547.
`[CCG00] F. Cali, M. Conti and E. Gregori, “Dynamic Tuning
`of the IEEE 802.11 Protocol to Achieve a Theoretical
`IEEE/ACM
`Trans.
`on
`Throughput Limit,”
`Networking, Vol. 8, Dec. 2000, pp. 785-799.
`[CLMK01] M.V. Clark, K.K. Leung, B. McNair and Z.
`Kostic, “Outdoor IEEE 802.11 Cellular Networks:
`Radio Link Performance,” Proc. IEEE ICC 2002.
`[CQW99] J. Chuang, X. Qiu and J. Whitehead, “Data
`Throughput Enhancement
`in Wireless Packet
`Systems by
`Improved Link Adaptation with
`Application to the EDGE System,” Proc. of IEEE
`VTC’99, Sept. 1999.
`[HT00] H. Holma and A. Toskala (Ed.), WCDMA for UMTS,
`John Wiley & Sons, New York, 2000.
`[I97] IEEE 802.11, “Wireless LAN Medium Access Control
`(MAC) and Physical Layer (PHY) Specification,”
`1997.
`[I99a] IEEE 802.11a, “Part 11: Wireless LAN Medium Access
`Control
`(MAC)
`and Physical Layer
`(PHY)
`Specification: High-Speed Physical Layer Extension
`in the 5 GHz Band,” 1999.
`[I99b] IEEE 802.11b, “Part 11: Wireless LAN Medium
`Access Control (MAC) and Physical Layer (PHY)
`Specification: High-Speed Physical Layer Extension
`in the 2.4 GHz Band,” 1999.
`[K76] L. Kleinrock, Queueing Systems, Vol. 2: Computer
`Applications, John Wiley & Sons, 1976, p. 398.
`[OP99] B. O’Hara and A. Petrick, IEEE 802.11 Handbook,
`IEEE Press, New York, 1999.
`[SAE98] P. Schramm, et. al., “Radio Interference Performance
`of EDGE, a Proposal for Enhanced Data Rates in
`Existing Digital Cellular Systems,” Proc. of IEEE
`VTC, Ottawa, Canada, May 1998, pp. 1064-1068.
`[VAM99] R. van Nee, et. al., “New High-Rate Wireless LAN
`Standard,” IEEE Commun. Mag., Dec. 1999, pp. 82-
`88.
`[VCM01] M. Veeraraghavan, N. Cocker and T. Moors,
`“Support of Voice Services in IEEE 802.11 Wireless
`LANs,” Proc. of IEEE Infocom 2001, Anchorage,
`Alaska, April 2001, pp. 488-497.
`
`IPR2020-00202
`Apple Inc. EX1059 Page 5
`
`