PRMA/DA: A New Media Access Control Protocol for Wireless ATM
`
`Indra Widjaja
`Jeong Geun Kim and
`Department of Electrical and Computer Engineering
`University of Arizona
`Tucson, AZ 85721
`e-mail: { j kkim, widj a j a}@ece. arizona.edu
`
`Abstract
`In this paper, we propose a media access control (MAC) pro-
`tocol f o r wireless local area networks (LANs) that is capable of
`supporting various types of t r a f l c demands, such as constant bit-
`rate (CBR) voice, variable bit-rate (VBR) video, and packet data.
`I n addition, the proposed protocol provides a seamless connectiv-
`ity t o a broadband ATM backbone network. O u r protocol, having
`a n air interface comparable to ATM, adopts a dynamic channel
`allocation scheme which enables expeditious network access and
`utilizes bandwidth resource ejjiciently. The simulation results pre-
`sented i n this paper shows the improvements of dynamic channel
`allocation over the static channel allocation scheme i n terms of
`k e y performance metrics such as: throughput, call blocking prob-
`ability, network access delay, and cell transmission delay.
`Introduction
`1
`With the rapid proliferation of high performance portable
`computers and mobile devices such a.s personal digital assistants
`(PDAs), the demand to connect all these devices to the fixed net-
`work (such as Asynchronous Transfer Mode) in a seamless fashion
`has been rapidly increasing. In particular, wireless local area net-
`works (LANs) are expected to be a crucial enabling technology in
`traditional office settings [l], [5]. A major technical issue related
`to wireless LANs is concerned with the selection of a suitable me-
`dia access control (MAC) protocol to efficiently arbitrate multiple
`mobile stations (MSs) on a common shared medium [ 5 ] . Several
`MAC protocols have been proposed for the wireless LANs. The
`protocols include Packet Reservation Multiple Access (PRMA),
`Code Division Multiple Access (CDMA), variants of PRMA, etc.
`PRMA is a statistical multiplexing method for delivering
`speech signals via a TDMA system [a], [7], [8]. Although PRMA
`is initially designed to carry voice traffic only, current efforts in
`computer networking for both wired and wireless segments are
`directed at supporting multimedia traffic [l]. Thus, future wire-
`less personal devices are expected to support multimedia applica-
`tions such as image retrieval, video conferencing as well as tradi-
`tional voice and data. To cope with the necessity of transporting
`other media plus voice, the enhanced version of PRMA called
`IPRMA (Integrated PRMA) is proposed to transmit both speech
`and data packets with higher throughput [4]. The novelty of the
`proposal comes from the fact that data users may reserve rriul-
`tiple slots across a frame to increase system throughput. While
`IPRM.4 is not intended to support multimedia traffic, the work
`by Raychaudhuri e t al. [l] introduces the multimedia capable inte-
`grated services wireless networks in an ATM-based t,ransport, ar-
`chitecture. In this study, multiservice-dynamic-reservation
`time-
`division-multiple-access (MDR TDMA) is employed as a MAC
`method. In the MDR 'TDMA, a TDMA frame consists of re-
`quest slots and message slots. The request slot is relatively small
`and is used for random access to the system based on the slotted
`ALOHA. The message slot is used to carry CBR (Constant Bit
`Rate), VBR (Variable Bit Rate), or ABR (Available Bit Rate)
`traffic once reservation is granted. Rather than using the fixed
`frame scheme, another enhanced variation of PRMA. called C-
`
`PRMA (Centralized PRMA), adopts a hybrid random access and
`polling scheme [3]. Random access techniques are used by the
`mobile stations to reserve slots for a message transmission, while
`the information packets are transmitted by using a polling scheme
`managed by the BS. The main features of C-PRMA are the capa-
`bility to integrate several classes of service and provide a prompt
`retransmission method for corrupted packets.
`In this paper, we propose an efficient MAC protocol for wire-
`less ATM, which is suitable for multimedia applications requiring
`a wide range of telecommunication services such as CBR, VBR,
`packet data (ABR), and others. Each application category has
`its own service requirements. Thus we need a MAC protocol that
`can meet the different requirements simultaneously in order to
`enable integrated transmission in a wireless ATM environment.
`Our proposal is based on the PRMA protocol and does not re-
`quire mini (request) slots which incur a certain fixed overhead. In
`particular, our proposed method increases the role of the BS to
`also act as a bandwidth allocation manager. A mobile station can
`get admitted to the network or is allowed to increase its trans-
`mission rate only when it gets permission from the BS. Mobile
`stations becoming active try to access the network by grabbing
`the available (unused) slot in a frame. The simultaneous access to
`the slot among the multiple stations leads to collision. To relieve
`the network access delay, we introduce the concept of dynamic
`allocation of available slots which may allow contending stations
`to have freer access to the available slots. The available slots can
`expand or shrink depending on the congestion state of the net-
`work. In addition, our protocol can flexibly allocate slots based
`on the traffic parameters specified by the users during the connec-
`tion set-up phase, such as the average number of cells generated
`during a frame period. The traffic parameters have an important
`role in deciding call admission.
`The structure of the paper is as follows. In Section 2 , we pro-
`pose a modified version of PRMA to handle multiservice wireless
`applications. In Section 3 , we describe the simulation model with
`pertinent numerical parameters. The simulation results are dis-
`cussed in terms of the important performance measures such as
`throughput, call blocking probability, network access delay, and
`cell transmission delay. We conclude the paper in Section 4.
`2 The Protocol
`2.1 Basic Operation
`The proposed MAC protocol, called PRMA/DA (Packet
`Reservation Multiple Access/Dynamic Allocation), is designed to
`operate in the conventional cellular system architecture. In the
`cellular architecture, a base station (BS) is located in the center
`of a cell with mobile stations (MSs) dispersed inside the cell. All
`the communication services present in the cellular architecture
`are made by having the BS relay the traffic between the commu-
`nication participants. Depending on the direction of transmis-
`sion between the MS and the BS, the communication channel in
`PRMA/DA is categorized into two separate time-slotted channels:
`uplink channel and downlznk channel. The uplink channel deliv-
`ers the information from the MSs to the BS, while the downlink
`
`0-7803-3250-4/96$5.000 1996 IEEE
`
`240
`
`Ericsson Exhibit 1009
`Page 1
`
`

`

`PRMA/DA
`HEADER
`
`ATM CELL
`HEADER
`
`48 BYTE
`PAYLOAD
`
`PRMNDA
`TRAILER
`
`SYNC
`
`NS
`
`other control fields
`
`SYNC :synchronization bits
`
`Figure 1: Cell format at the air interface.
`
`channel is used to communicate in the opposite direction. For the
`uplink transmission, the mobile stat,ions in a cell share the com-
`munication channel using the PRMA/DA protocol. The downlink
`channel operates with a contention-free TDM (time-division mul-
`tiplexing) broadcast mode.
`PRMA/DA adopts a packet switching scheme comparable to
`the ATM transport architecture. Next generation wireless local
`networks will be required to co-exist with ATM network, which
`should be deployed in the near future [l]. Based on the consid-
`erations, we adopts the ATM cell relay paradigm as the basic
`transport architecture. Thus an ATM cell acts as a fundamen-
`tal unit of protocol processing and switching in both wired and
`wireless network segments. The typical structure of the transport
`cell at the air interface is illustrated in Fig. 1. As shown in the
`figure, the ATM cell is encapsulated by a PRMA/DA header and
`trailer. The header contains synchronization bits, a NS field indi-
`cating the number of slots requested from a mobile station, and
`other control fields, while the trailer minimally contains an error
`check field. The NS field is of importance to the operation of the
`protocol, since the field is used to carry the information on the
`current bandwidth demand of a mobile station to the BS at every
`frame.
`PRMA/DA operates on the frame basis.
`Time on the up-
`link channel is divided int,o a contiguous sequence of PRMA/DA
`frames, which are further subdivided into a fixed number of slots.
`In particular, a PRMA/DA frame can be segmented into avail-
`ableslots and reservationslots consisting of CBR reservation slots,
`VBR reservation slots, and ABR reservation slots as illustrated
`in Fig. 2. The BS has absolute control in determining the num-
`ber of both available slots and reservation slots. Furthermore,
`the BS also specifies the number of slots assigned to each individ-
`ual reserving station. The number of available slots depends on
`the intensity of demand to access the network among the mobile
`stations. In contrast, the number of reservation slots assigned to
`each reserving station is primarily dependent on the statistical
`properties of traffic a MS intends to transmit.
`Each group of slots in a frame provides specific service charac-
`teristics. The available slots provide a communication mechanism
`for a mobile station to attain network access, while t,he reserva-
`tion slots supply a mobile station with a major resource of channel
`bandwidth during a network access (or connection) period. In or-
`der to transport the traffic, an activated station needs to access
`the network. The network access is made by transmitting a cell
`in a randomly selected available slot. After completing the con-
`tention procedure, the mobile station can use the reservation slots
`without undergoing further contentions.
`The contention procedure of the PRMA/DA protocol oper-
`ates on a random access (slotted ALOHA) scheme. A mobile
`station, when it has just become active, switches its mode into
`the contentaonmode from the inactivemode. A mobile station in
`the contention mode is called a contendingstation in PRMA/DA
`terms. As in the slotted ALOHA, a contending station randomly
`selects one of the available slots which occupy the beginning por-
`tion of a frame. The available slots are numbered from 1 to N ,
`as shown in Fig. 2. A contending station randomly chooses a
`number ranging from 1 to the number of available slots (N,) and
`transmits the cell in the selected slot. As an example, the mobile
`
`-I
`
`l-
`
`-I
`
`I--
`
`/
`
`/
`I
`
`vanable
`
`+ +
`
`vanable
`
`variable
`
`+
`
`'\
`
`\
`I
`
`slots
`
`slots
`
`slots
`
`slots
`
`PRMNDA FRAME
`
`Figure 2: The PRMA/DA fram.e format
`
`station that chooses the random number k (1 5 k 5 N,), must
`transmit its first cell in the k th available slot. Right after the
`contention period, the BS advises the contending stations whether
`the network access is successful or not.
`Unless other contending stations try to transmit their cells in
`the same slot, the mobile station will attain the network access
`and shift its mode to the reservation mode. If the mobile station
`fails to acquire the network access due 'to a. collision, i t needs to
`repeat the contention procedure at the next frame. The BS broad-
`casts the new N , value at the end of each frame. Upon receiv-
`ing the information, the contending stations prepaxe for the next
`contention period. For CBR/VBR traffic, which has stringent
`timing constraints, the unbounded repetition of the contention
`procedure is meaningless. Therefore, we specify the maximum
`time for contention which is called maximvm setup time (WmaZ).
`If a contention period of a mobile statLon lasts longer than the
`maximum setup time, its call will be discarded and the mobile
`station will return to the inactive mode.
`The number of the available slots in PRMA/'DA varies dy-
`namically depending on the congestion state of thie network. The
`objective of the scheme is to maximize the throughput by having
`the majority of slots serve the reserving stations while maintain-
`ing a minimal number of available slots for contending mobile
`stations. As the demand for network access increases, the num-
`ber of available slots will expand according to the aJgorithm which
`will be discussed later in detail. With decreasing requests for net-
`work access, the number of available slots will shrink accordingly.
`Finally the number of available slots reduces to one when no de-
`mand exists.
`The number of reservation slots assigned to the reserving sta-
`tions is governed by the dynamic allocation algorithm. The pri-
`mary factor considered in the algorithm is the statistical proper-
`ties of the traffic that a mobile station intends to transport. At
`the connection setup phase, a contending station is required to
`specify the traffic parameters in the cell that is delegated to con-
`tend for an available slot. In the current phase of our work, the
`statistical properties of the traffic are representemd only in terms
`of the average number (R,) and peak number (&) of cells gener-
`ated during a frame period. Further extension will be required to
`more accurately characterize the traffic. Along with the parame-
`ter ( R m ) , the current bandwidth demand which is delivered by the
`N S field is considered in determining the number of reservation
`slots assigned to a reserving station. .4t the end of each frame,
`the BS broadcasts the information about the number of reserva-
`tion slots which have been assigned to each reserving station and
`the respective slot locations. Upon receiving the information, the
`reserving stations can begin transmission in their assigned slots
`in a frame.
`
`24 1
`
`Ericsson Exhibit 1009
`Page 2
`
`

`

`Dynamic Allocation of Reservation Slots
`2.2
`In a mixed traffic environment, each category of traffic, e.g.,
`CBR, VBR, and ABR, has its own unique traffic properties and
`service requirements to maintain the declared QoS (Quality of
`Service). For instance, CBR and VBR streams have a stringent
`timing constraint. Furthermore, the bandwidth demand of VBR
`and ABR traffic fluctuates due to their bursty nature, while CBR
`traffic claims constant quantity of bandwidth. To cope with the
`heterogeneous and varying requirements, dynamic channel allo-
`cation is chosen rather than a static channel allocation which
`might cause inefficiency. In order to enable the dynamic alloca-
`tion scheme to operate efficiently, the BS requires the following
`parameters: the average traffic rate (&),
`the peak rate ( R p ) ,
`and the number of slots requested by a reserving station which is
`carried in the NS field of a cell. These parameters are the major
`factors considered by the BS when determining the number of
`slots to assign to a reserving station.
`Once the BS determines the number of available slots (N,),
`the rest of the slots in a frame are used to serve the reserving
`stations as shown in equation (1).
`
`where Nf is the total number of slots in a frame and N , is the
`total number of reservation slots in a frame.
`The total number of reservation slots (N,) is distributed to
`each class of traffic based on the priority of the traffic.
`In
`PRMA/DA, three categories of traffic, i.e., CBR/VBR/ABR, are
`prioritized in terms of their respective timing constraints. Thus
`the CBR/VBR reserving station has a priority in service over the
`ABR station. First, every CBR reserving station, z, takes Rp,t
`slots since it has constant bandwidth demand which is equivalent
`to its traffic parameter Rp,t. The total number of slots assigned
`to all CBR reserving stations, N r
`, ~ ~ ~
`is
`
`where Sc is the set of CBR reserving stations. Second, the
`total number of slots assigned to VBR traffic must be at least
`E,,,, Rm,J, where Sv is the set of VBR reserving stations. The
`- ~,,,, leftover (N, - N T , c ~ ~
`is assigned to the ABR
`
`stations. If the 4 B R stations do not consume all of the available
`slots, the remainder is reallocated to the VBR reserving stations.
`It is worthwhile to transfer unused slots to a station with higher
`bandwidth demand. This idea is reflected in equation (3):
`
`is the total number of VBR and ABR
`where N?,VBR and N p , ~ ~ ~
`reservation slots respectively, and Dv and D D is total number of
`slots requested by the reserving VBR and ABR stations respec-
`tively.
`Slots can now be allocated to individual reserving stations.
`The policy for the slot allocation to an individual reserving station
`is very similar in principle to the case of the slot allocation to each
`group. That is, the reservation slots of each group are distributed
`in proportion to the average traffic rate (the parameter R,) of an
`individual station. As in the case of slot allocation for each traffic
`group, the surplus slots gathered from the stations which request
`less will be distributed to the stations which request more, again
`based on their average traffic rate ( E m ) .
`
`<1.0,0> <0,0.1> n
`
`<0,2,0>
`
`Figure 3: The state transition diagram for the number of available
`slots
`
`2.3 Dynamic Allocation of Available Slots
`The objective of dynamic allocation of available slots is to pro-
`vide the maximum achievable bandwidth to the reserving stations
`and to minimize the network access delay of contending stations.
`In the PRMA/DA, the BS maintains only a single available slot
`when no demand for network access exists. However as the traffic
`load increases, the single slot is not sufficient to handle the net-
`work access demand of multiple contending stations. Therefore,
`the number of available slots needs to adapt to the intensity of
`the network access demand.
`In the PRMA/DA, the number of available slots ( N , ) is ad-
`justed such that it approximately corresponds to the number of
`contending stations. The BS cannot know the exact number of
`contending stations present in a cell, although it can estimate the
`number of contending stations which were involved in collisions
`in the previous frame. The BS can estimate the least number of
`contending stations by observing the result of network access at
`the previous frame.
`As an example, suppose that there exists a single available slot
`with no access demand. Also suppose that, at a certain frame,
`a collision occurs at the slot. Since the BS can only infer that
`at least two contending stations are involved in the collision, it
`expands the available slots to two. At the next frame, if collisions
`occur at the two available slots, it can be estimated in the same
`way that at least four contending stations are present. The BS
`then increases the available slots to four. At the following frame,
`if there exist two successful accesses, one collision and one unused
`slot out of four available slots, the BS decreases the available
`slots to two since at least two contending stations still request the
`network access. As explained in this example, the BS keeps track
`of the minimum number of contending stations. The number of
`contending stations can be estimated from the number of available
`slots ( N a ) , the number of successful access slots ( N s ) , t,he number
`of collision slots ( N c ) , and the number of unused slots (N,) at
`the previous frame, where Nu = N , + N , + Nu. The procedure
`to determine N , is described by the state transition diagram in
`Fig. 3 . As shown in Fig. 3, the state is represented by N , and the
`state change is triggered by (Nsl N,, N u ) .
`The dynamic allocation for the available slots can be formal-
`ized as follows:
`
`242
`
`Ericsson Exhibit 1009
`Page 3
`
`

`

`2
`f
`
`where A@) is the number of available slots at the (k) th frame,
`N i k ) is the number of slots in which collision occurs, and N i k ) is
`the number of slots in which successful network access is made.
`'In the PRMA/DA, CBR/VBR reserving station i is assured to
`receive at least R,,; slots which corresponds to minimum band-
`width for maintaining t,he declared &OS. Therefore, in order to
`guarantee the minimum bandwidth for the &OS, the BS exercises
`the connection admission control (CAC) to prevent the admission
`of a new call from degrading the QoS of the reserving stations.
`certain maximum ( N j - ciEs,
`Rp,i - ~,,,, Rm,J) as shown in
`Along with CAC, the number of available slots is bounded by a
`equation (4) to protect the CBR/VBR reserving stations.
`3 Simulation and Performance
`In this section, we provide numerical performance results for
`the PRMA/DA protocol operating in a mixed voice/video/data
`environment and compare
`it with
`its counterpart called
`PRMA/FA (Fixed Allocation). PRMA/FA has the exactly same
`principle of operation as PRMA/DA except that it does not ex-
`ercise dynamic allocation of the available slots. That is, rather
`than delegating the BS to control the number of available slots
`according to the intensity of network access demand as in the
`PRMA/DA, PRMA/FA provides the available slots, if available,
`after assigning slots to the reserving stations. Therefore, the net-
`work access is dependent merely on bandwidth demands of the
`reserving stations. In the case where the bandwidth demand of all
`reserving stations exceeds the total channel capacity, a contend-
`ing station will be forced to wait until an available slot is released
`by the reserving stations.
`We use discrete event simulation to characterize the perfor-
`mance of the PRMA/DA protocol in a multiservice environment.
`The communication channel is partitioned into frames and each
`frame contains a fixed number of slots ( N f = 100). The simu-
`lation is based on a channel speed of approximately 7.067 Mb/s
`which corresponds to the capability of transporting, per one frame
`period ( T ) , one hundred voice ATM cells digitized by a 64 Kb/s
`codec. In the current phase of work, we do not include the use
`of the VAD (voice activity detector). Thus the CBR stream (64
`Kb/s) is generated during the voice call duration.
`The users accessing the uplink channel in each cell are classified
`into three sets of users: CBR, VRR, and ABR users. A new
`and the call duration (T,)
`voice call arrives at the rate of A,
`is exponentially distributed with an average of 3 minutes. We
`fix the number of VBR traffic sources to five which arrive at the
`beginning of the simulation. The duration of the VBR connection
`covers the entire simulation. A ABR call arrives at the rate of
`Ad and the length of the packet is exponentially distributed with
`average L (5.12 Kbytes) which corresponds to about 107 cells. In
`order to model a VBR video traffic source, we use the video codec
`model proposed by Heyman e t al. [GI. In the model, the number
`of video cells per frame is determined by a gamma distribution (or
`equivalently negative binomial) and a DAR (1) model [GI.
`It is assumed that a mobile station has infinite buffer capac-
`ity. Also an ideal communication channel i s assumed, implying
`that transmission errors and retransmissions do not occur in the
`simulation. The simulation parameters are summarized in Table
`1.
`3.1 Numerical Results
`The numerical results are presented here to evaluate the suit-
`ability of the proposed MAC protocol in the multiservice envi-
`ronment.
`In the experiments, the performance of two channel
`
`Figure 4: Throughput versus voice call offered traffic
`
`allocation schemes-dynamic and fixed, is evaluated in terms of
`key measures. The measures considered here include throughput,
`transmission delay, network access delay, and call blocking. The
`simulation model was exercised over a wide range of CBR traffic
`loading (0.0 N 0.1908) on both a PRMA/.DA (dynamic) and a
`PRMA/FA (fixed) channel. All the curves in the following are
`plotted with 95% confidence interval.
`Figure 4 shows the curves of throughput versus CBR offered
`traffic for tlhe two compared cases, given that the ABR call arrival
`rate X d = 0.05 (calls/sec). As expected, the throughput curves are
`initially quite linear, and eventually saturate at the high load. At
`light offered traffic, the two schemes do :not exhibit any apparent
`difference in throughput. However, as the CBR offered traffic
`increases, the curves start to show wider gaps in throughput. The
`result on throughput is closely related to the difference in CBR
`call handling between the two compared schemes.
`As discussed earlier, the PRMA/DA reserves at least a single
`available slot for a new network user and the number of available
`slots is adjusted depending on the intensity of the network access
`demand. In contrast, PRMA/FA does not control the available
`slots. All the slots are used by the reserving st,ations as long
`as their bandwidth demand exceeds the channel capacity (to-
`tal number of slots N f ) . Thus, available slots are created only
`when the total number of slots requested by all reserving stations
`goes below the total number of slots in a frame ( N f ) . Thus, the
`outcome of an activated station attempting network access de-
`pends heavily on the current bandwidth demand of the reserving
`stations. However, in PRMA/DA, a contending station will get
`admitted to the network unless its admission degrades the per-
`formance of reserving stations, even when total n.umber of slots
`requested by the reserving stations exceeds the total number of
`slots ( N j ) . The difference in the allocation scheme of the two
`protocols does not significantly lead to performance degradation
`at light load conditions (0.0 N 0.07) as seen in Fig. 4. At heavy
`loads, PRMA/FA is more likely to experience call blocking, thus
`causing lower throughput.
`Curves for voice blocking probability versus CBR offered traffic
`are shown in Fig. 5 for two compared channel allocation schemes.
`The blocking probability using PRMAIFA starts to go up steeply
`at the point of offered traffic 0.06, while PRMA/DA at 0.12 ap-
`proximately. From Fig. 5, it is observed that PRMA/FA has non-
`negligible blocking probability in the light loading range (0.02 N
`0.06). This is expected since network access in PRMA/FA mainly
`depends on the current slot demand of t8he reserving stations.
`In this simulation, the VBR video source generates traffic corre-
`sponding to 16 cells per frame on average (R, = 16). Thus total
`video traffic accounts for around 80% of total traffic. During the
`simulation, the video sources are observed to temporarily gener-
`ate traffic at the rate of more than 100 dots per frame since their
`traffic fluctuates severely. Thus if a mobile station tries to att,ain
`network access during the period when all the slots serve the re-
`serving stations and the period lasts longer than the maximum
`call setup time (W,,,,),
`the call will be blocked.
`
`243
`
`Ericsson Exhibit 1009
`Page 4
`
`

`

`I item
`I Frame lenrth (msec)
`Slot length (msec)
`
`1 symbol I
`I
`T
`1
`I
`
`value
` 6
`0.06
`
`1
`I
`
`\
`
`/ I
`
`\
`
`I
`
`/
`
`
`
`PRMA/DA channel speed (Kbps)
`Voice codec data rate (KbDs’I
`CBR/VBR/ABR slot size (bytes)
`Payload of a single slot (bytes)
`Arrival rate (exDonential’I of new voice calls icalls/sec/user)
`Average length (exponential) of a voice call (min)
`Maximum (fixed) voice call set up time(sec)
`Video IVBR’I frame rate
`Average number of cells per video frame
`
`\
`
`-
`
`
`
`,
`
`,
`
`I
`
`,
`
`I
`
`R
`J
`I R,,
`
`
`
`A,,
`
`T C
`
`WO,,,
`
`~~
`
`I
`
`I
`I
`
`I
`
` 7067
`64
`. -
`53
`48
`0.01 N 0.1
`3
`5
`25
`I
`104.8
`I 0.01 N 0.1
`I
`5.12
`
`arrival rate (exponential) of new ABR calls (calls/sec/user)
`Average message length (exponential) of ABR calls (Kbytes)
`
`A d
`L
`
`Figure 5 : Voice call blocking probability
`versus voice call offered traffic
`
`Figure 6: Voice call network access de-
`lay versus voice call offered traffic
`
`Figure 7: VBR video cell transmission
`delay versus voice call offered traffic
`
`Figure 6 shows the curves of CBR network access delay versus
`CBR offered traffic for the two channel allocation schemes. The
`curves have a very similar shape when compared to the curves
`for blocking probability. That is, PRMA/FA shows relatively
`longer delay even at light load while PRMA/DA starts to undergo
`noticeable delay when the offered traffic reaches 0.13.
`Figure 7 shows the VBR video cell transmission delay versus
`the offered CBR traffic. As shown in Fig. 7, a video cell experi-
`ences longer delay as higher loads are applied. The video cell of
`PRMA/F’A has a gradual increase in delay while the PRMA/DA
`has a steep increase. Generally, as the CBR offered traffic in-
`creases, the PRMA/DA has more CBR reserving stations than
`PRMA/FA does due to PRMA/DA’s advantageous way of ac-
`cepting new calls. As more CBR stations enter the network, VBR
`reserving stations will lose more slots since CBR stations have a
`higher service priority. The performance of VBR stations can be
`improved if proper call admission control is applied.
`
`4 Conclusion
`In this paper, we have presented a new MAC protocol called
`PRMA/DA for wireless ATM LANs. PRR;IA/DA adopts a dy-
`namic channel allocation scheme to cope with varying bandwidth
`demands for several classes of service, e.g., CBR, VBR and ABR.
`Furthermore, the ATM cell is employed as a basic transmission
`unit to provide seamless connectivity to a broadband ATM back-
`bone network.
`traffic environment.
`The results show that for a mixed
`PRMA/DA does achieve a significantly higher bandwidt 11 effi-
`ciency over the static channel allocation scheme. The perfor-
`mance results also demonstrate that PRMA/DA provides expe-
`ditious network access and less blocking probability for the CBR
`user
`
`References
`[I] D. Raychaudhuri and N. D. Wilson, “ATM-based transport
`architecture for multiservices wireless personal communica-
`tion networks,” J . Select. Areas Commun., vol. 12, no. 8,
`pp. 1401-1414, Oct. 1994.
`[2] D. J. Goodman, “Cellular packet communications,” IEEE
`Trans. Commun., vol. 38, no. 8, pp. 1272-1280, Aug. 1990.
`[3] G. Bianchi, F. Borgonovo, L. Fratta, L. Musumeci and M.
`Zorzi, “C-PRMA: the centralized packet reservation multiple
`access for local wireless communications,” GLOBECOM ’94,
`San Francisco, USA, pp. 1340-1346, Nov. 1994.
`[4] W. Wong and D. J . Goodman, “Integrated data and speech
`transmission using packet reservation multiple access,” ICC
`’93, Geneva, Switzerland, pp. 172-176, 1993.
`[5] K . C. Chen, “Medium access control of wireless LANs for
`mobile computing,” IEEE Network, pp. 50-63, Sep. 1994.
`[6] D. IIPyman, T. V. Lakshman, A . Tabatabai and H. Heeke,
`“Modeling teleconference traffic from VBR video coders,”
`SD-PERCOMM/ICC ’94, New Orleans, USA, pp. 1744-1748,
`May 1994.
`[7] D. J . Goodman, R. A. Valenzuela, K . T. Gayliard and B.
`Ramamurthi, “Packet reservation multiple access for local
`wireless communication,” IEEE Trans. Commun., vol. 37,
`no. 8. pp. 885-890. Aug. 1989.
`[8] S. Nanda, D. J. Goodman and U. Timor, “Performance of
`PRMA: A packet voice protocol for cellular systems,” IEEE
`Trans. Vehic. Technol.. vol. 40, no. 3 , pp. 584-598, Aug.
`1991.
`
`244
`
`Ericsson Exhibit 1009
`Page 5
`
`