Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e: challenges and issues
DDate of publication xxxx 00, 0000, date of current version xxxx 00, 0000.
Digital Object Identifier 00.0000/ACCESS.2018.DOI
Review on QoS provisioning approachesfor supporting video traffic inIEEE802.11e: challenges and issues
MOHAMMED A. AL-MAQRI , MOHAMED A. ALRSHAH ,(Senior Member, IEEE), ANDMOHAMED OTHMAN. ,(Senior Member, IEEE) Department of Communications Technology and Networks, Universiti Putra Malaysia, Serdang, 43400, Malaysia Azal University for Human Development, 60th Street, Sana’a, Yemen Computational Science and Mathematical Physics Lab, INSPEM, Universiti Putra Malaysia, Serdang, 43400, Malaysia
Corresponding author: Mohamed A. Alrshah (e-mail: [email protected])This research was funded by the Malaysian Ministry of Education under UPM/700-2/1/GPB/2017/9557900 Putra with High-Impact Grant.
ABSTRACT
Recently, the demand for multimedia applications is dramatically increased, which in turnincreases the portion of video traffic on the Internet. The video streams, which require stringent Qualityof Service (QoS), are expected to occupy more than two-thirds of web traffic by 2019. IEEE802.11ehas introduced HCF Controlled Channel Access (HCCA) to provide QoS for delay-sensitive applicationsincluding highly compressed video streams. However, IEEE802.11e performance is hindered by thedynamic nature of Variable Bit Rate (VBR) video streams in which packet size and interval time arerapidly fluctuating during the traffic lifetime. In order to make IEEE802.11e able to accommodate withthe irregularity of VBR video traffic, many approaches have been used in the literature. In this article, wehighlight and discuss the QoS challenges in IEEE802.11e. Then, we classify the existing QoS approaches inIEEE802.11e and we also discuss the selection of recent promising and interesting enhancements of HCCA.Eventually, a set of open research issues and potential future directions is presented.
INDEX TERMS
I. INTRODUCTION T HE optimal transport of delay-constrained multimediaservices over WLANs requires adaptation to manyaspects of Open Systems Interconnection (OSI) modellayers starting from delay constraints and bandwidthvariations of the traffic at the application layer up toaccommodation to wireless channel conditions and powerconstraints at the physical layer. The efficiency of 802.11eHCF Controlled Channel Access (HCCA) function mainlydepends on the accuracy of its scheduler in assigningnetwork resources, such as channel bandwidth, to the trafficstreams without jeopardizing the QoS constraints such asdelay and throughput. Moreover, with the presence ofdelay-sensitive multimedia traffic with variable profile, theexisting scheduling approaches become inefficient. Thus, thescheduler is required to consider the fluctuation of traffic inthe scheduling process.This article introduces an overview of the prime challengesfor provisioning QoS for multimedia traffic with emphasizeon Variable Bit Rate (VBR) traffic in IEEE 802.11e wireless networks. Then, it presents a taxonomy for the existingsolutions, and describes the most representative properties,advantages, and design challenges. This taxonomy comprisesthe core approaches and techniques on IEEE802.11eprotocol, with more emphasize on HCCA enhancements.Additionally, a systematic summarization and comparisonfor research contributions in each field are used to clearlyidentify the current challenges for further research. Finally,the article discusses the most critical issues which hinderthe provisioning of QoS in wireless networks with a specialattention to polling and Transmission Opportunity (TXOP)allocation enhancements.This paper is a survey of QoS provisioning for videotransmission in IEEE802.11e, which is organized as follows:Section II exhibits the background about IEEE802.11estandard and its functions. Section III presents the mainchallenges in IEEE802.11e WLANs. Section IV classifiesand reviews the core approaches in IEEE802.11e WLANs,which were proposed to enhance QoS provisioning formultimedia traffic. A number of leading approaches aiming at
VOLUME 0, 2018 a r X i v : . [ c s . N I] F e b . A. Al-Maqri et al. : Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e improving the QoS for multimedia traffic has been discussedin Section V. Section VI shows a general comparison ofthe IEEE802.11e approaches and their targeted features, andlists some of the strength and limitation criteria of theseapproaches. Section VII and Section IX identifies researchtrends, challenges, and potential future areas related to thearticle’s scope, and finally Section X concludes the article. II. IEEE802.11E STANDARD
Several amendments have been made to the legacyIEEE802.11 WLAN standard [1], as shown in Table 1.IEEE802.11e is one of the approved versions of IEEE802.11standard, which defines a combination of Quality of Service(QoS) improvements on the Medium Access Control (MAC)layer for WLAN applications, as shown in Fig. 1. Thestandard is critically important for applications that are verysensitive to delay, such as Voice over WLAN (VoWLAN) andmultimedia streaming.In IEEE802.11e, the QoS feature includes an extracoordination function called Hybrid Coordination Function(HCF). This function combines both functionalities ofthe well-known Point Coordination Function (PCF) andDistributed Coordination Function (DCF). In order topermit the use of a uniform assortment of frame exchangesequences for QoS data transfers during the time of bothContention Period (CP) and Contention Free Period (CFP),the HCF introduced some enhanced frame subtypes andQoS-specific mechanisms. As for contention-based transfer,HCF employs a contention-based channel access approach,namely Enhanced Distributed Channel Access (EDCA),while for contention-free transfer it uses a controlled-channelaccess method, so-called HCCA. Stations (STAs) mightobtain TXOPs using EDCA, HCCA or both schemestogether. Thus, a TXOP is defined as EDCA TXOP if it isobtained by the contention-based channel access, while it isdefined as HCCA-TXOP if it is obtained by the controlledchannel access.
FIGURE 1.
MAC architecture in IEEE802.11e.
A. ENHANCED DISTRIBUTED CHANNEL ACCESS(EDCA)
EDCA mechanism has been designed to provide sort ofdifferentiated distributed access to Wireless Medium (WM) for STAs by using eight uneven User Prioritiess (UPs). Itdetermines four Access Categories (ACs) to provide supportfor traffic delivery at the STAs using UPs, which produces theAC, as shown in Table 2. For every AC, an enhanced variantof DCF, called Enhanced DCF (EDCF), contends for TXOPsusing a set of EDCA parameters. For more details about theEDCF refer to [11]. Implementation of this mechanism iseasy; however, the QoS requirement of a realtime traffic cannot always be met, especially when the heavy load conditionsoccur. In heavy loaded scenarios, higher prioritized trafficQoS requirement may easily be broken even though itexhausts most of the available bandwidth. However, lowerprioritized traffic may be starved and severely deteriorated inboth efficiency and effectiveness.
B. HCF CONTROLLED CHANNEL ACCESS (HCCA)
As known in IEEE-802.11e, a synchronization signal isrhythmically sent to all of the connected stations in the BasicService Set (BSS). The time between two subsequent signalsmakes a super-frame, where a service can be deliveredthrough this super-frame over two periods of time, CFP andCP. The data of any station has to be transmitted duringa period of time, namely TXOP, which is dedicated for aQoS-enabled Station (QSTA) to transfer its MAC-ServiceData Units (MSDUs). Fundamentally, TXOP is acquiredthrough the contention-based access, which is known asEDCA-TXOP. As for the controlled medium access, theHybrid Coordinator (HC) grants the TXOP to the QSTA(known as polled TXOP). Fig. 2 shows a clear example of802.11e super-frame which demonstrates the interchangingof one controlled medium access and one contention-basedperiod, where the later includes one QoS-enabled AccesPoint (QAP) and three QSTAs. In general, controllingmedium access occurs either within the CP or through theCFP if the medium remains idle for at least one period ofPCF Inter Frame Space (PIFS). In order to support QoSin HCCA, many researchers have proposed to improve theexisting PCF by controlling the transmission only within theCFP. Therefore, the data packets of any wireless station inHCCA can be only transmitted during a declared period oftime in the poll frame.
FIGURE 2.
An 802.11e super-frame example, CFP and CP. In the CFP, theframe exchange takes a place throughout the polling mechanism, while in CPthe QSTAs have to listen to the medium transmitting data packets.2
VOLUME 0, 2018 . A. Al-Maqri et al. : Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e
TABLE 1.
The family of IEEE-802.11 versions
Standard Objective Frequency and ModulationIEEE802.11 [2] To provide up to 2 Mbps bit rate. 2.4 GHz by utilizing DSSS and FHSS.IEEE802.11a [3] To provide up to 54 Mbps bit rate. 5 GHz by utilizing OFDM.IEEE802.11b [4] To provide up to 11 Mbps bit rate. 2.4 GHz by utilizing HRDSSS.IEEE802.11c [5] To ensures proper bridging operations. -IEEE802.11d [6] To covers more regulatory domains. -IEEE802.11e [7] To define new QoS enhancements to 802.11a and 802.11b. -IEEE802.11f [8] To provide interoperability for roaming among different APs. -IEEE802.11g [9] To provide up to 54 Mbps bit rate. 2.4 GHz by utilizing OFDM.IEEE802.11n [10] To provide up to 600 Mbps bit rate. 2.4 and 5 GHz by utilizing MIMO-OFDM.
TABLE 2.
Mappings of user priority to access category
Priority UP AC DesignationLowest 1 AC_BK Background2 AC_BK Background0 AC_BE Best Effort3 AC_BE Best Effort4 AC_VI Video5 AC_VI Video6 AC_VO VoiceHighest 7 AC_VO Voice
1) Reference design of HCCA
At the point if a QSTA wants to transmit its realtime TrafficStream (TS) within the contention-free period, it has tosend an ADDTS-Request to the QAP. This ADDTS-Requestdeclares the requirements of QoS for that specific TSwithin the relevant TS Specification (TSPEC) domain.Consequently, the QAP will try to fulfill the requirementswhile conserving the QoS of existing admitted flows. Ifthe ADDTS-Request is accepted, the QAP will reply anADDTS-Response back to the relevant station, then, thisstation will be admitted to the QAP polling list. Table 3 showsthe compulsory TSPEC parameters and their symbols.
TABLE 3.
Symbols used for TSPEC and scheduling parameters
Notation Description ρ Mean Data Rate L Nominal MSDU Size M Maximum MSDU Size D Delay Bound SI Service Interval mSI
Minimum Service Interval
MSI
Maximum Service Interval R Physical Transmission Rate BI Beacon Interval O Physical Layer mode (PHY) and MAC Overhead N Number of packets T super-frame duration T CP Contention-based duration
After accepting new ADDTS-Request, the HCCAscheduler will go through the following steps:1)
Assigning service interval
HCCA [11] computes the Service Interval (SI) as asub-multiple of the whole Beacon Interval BI , whichis calculated as the minimum of the maximum SIs of allpriorly accepted traffic streams including the incoming data traffic. Equation (1) is used to calculate the SI: SI = BI (cid:108) BIMSI min (cid:109) , (1)where M SI min is computed as in Equation (2):
M SI min = min ( M SI i ) , i ∈ [1 , n ] , (2)where M SI i denotes the maximum SI of the i th stream and n denotes the number of all previouslyadmitted QSTAs’ traffic streams.2) Allocating TXOP
Variant TXOP is allocated by HC to every acceptedQSTA based on the declared QoS parameters in theTSPEC, which allows the QSTA to obtain the requiredQoS. The HC calculates TXOP for the i th QSTA basedon the expected MSDUs, which may arrive at ρ i , ascalculated in Equation (3): N i = (cid:24) SI × ρ i L i (cid:25) , (3)where L i denotes the MSDU of the i th station.Thereafter, the TXOP of the i th station ( T XOP i ) iscalculated as the required time to transmit N i MSDUor one maximum MSDU at the relevant physical rate R i , as in Equation (4) below: T XOP i = max (cid:18) N i × L i R i + O, MR i + O (cid:19) (4)where O represents the total overhead, includingMAC and physical headers, poll frames overheads,inter-frame spaces (IFSs) and acknowledgments.3) Admission control
The Admission Control Unit (ACU) regulates theadmission of the TS while maintaining the QoS of thepreviously admitted TSs. When the ACU receives arequest of admitting a new TS, the ACU calculates anew SI using Equation (1) and estimates the numberof MSDUs that may arrive at this new SI based onEquation (3). Then, the ACU calculates the T XOP i for the particular TS using Equation (4). Finally, theACU would admit the relevant TS only if the followinginequality is satisfied: T XOP n +1 SI + n (cid:88) i =1 T XOP i SI ≤ T − T CP T (5) VOLUME 0, 2018
3. A. Al-Maqri et al. : Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e
Fig. 3 shows an example of an admitted stream from
ST A i . The beacon interval is 100ms and the maximumSI for the stream is 60ms. The scheduler sets ascheduled SI to 50ms with complying to Equation (5),where n represents the number of all admitted streams, n + 1 denotes the index of incoming TS, T indicatesthe beacon interval and T CP is the time reserved forEDCA contention-period.The HC sends an ADDTS-Response to the relevantQSTA only if Equation (5) is satisfied, and it sends amessage of rejection otherwise. Then, the HC will addthe accepted TS to its polling list. FIGURE 3.
Schedule for streams from STAs i to k. The streams are scheduledin Round-Robin fashion govern by the admission control unit
FIGURE 4.
QoS architecture of the IP Network. The QoS parameters aredefined in the MAC layer
III. QOS CHALLENGES IN IEEE802.11E WLANS
QoS is the overall effect of the service performance, whichdefines the satisfaction degree of a service user and manifestsitself in a number of subjective or objective parameters[12]. There are two ways to investigate the QoS, subjective(perceptive) and objective (network) measurements. In thesubjective measurement, the user involves to carry out aseries of assessment tests, while in objective measurement,typical network performance throughput, packet loss, packetjitter and delay is evaluated. In order to meet the usersatisfaction, the subjective QoS parameters shall be translatedinto a set of objective QoS parameters, e.g. throughput, delayand losses.QoS could be supported in different ways at differentprotocol layers as illustrated in Fig. 4. Some applicationshave the capability to adapt the generated traffic to theconditions of the underlying network in order to meetuser expectations. An example is the use of the Real-time Transport Protocol (RTP) and associated RTP ControlProtocol (RTCP) [13] to dynamically adapt the parametersof an audio and/or video streams, minimizing the losses dueto congestion in the network [14]. Nevertheless, applicationlayer mechanisms are usually not enough, since end-to-endQoS requires support in the lower layers of the protocol stackthroughout the network nodes that the traffic must traversefrom sender to receiver. However, this work mainly concernswith QoS provisioning at MAC layers.The QoS provisioning of diverse multimedia streams ina wireless environment imposes a chain of challenges dueto many factors of OSI model layers [15], [16], [17]ranging from traffic characteristics in application layer downto the wireless channels nature in physical layer. In thissection, a review of the major challenges that may emergewhen providing QoS for delay-sensitive applications inIEEE802.11e wireless networks.
A. ADAPTATION TO FLUCTUATION OF APPLICATIONPROFILE
Generally, the application profile of a traffic is defined by thealternation of the traffic over the time. The QoS provision ofa VBR flow is substantially influenced by the variation of theapplication profile over the time. The accurate estimation ofthe traffic at the application layer can significantly enhancethe performance of underlying functions of MAC layer toadapt its parameters according to these changes.The VBR video source can be generally classified intothree main categories [18], [19]: I) variable packet size withconstant Generation Interval (GI), e.g., MPEG-4 videos;II) constant packet size with variable GI, e.g., Voice overInternet Protocol (VoIP); and III) variable packet size withvariable GI, e.g., H.263.The transmission of video streams can be significantlyaffected by the compression techniques used, such asMPEG-4 and H.263. The nature of the frame structure andthe compression algorithm used along with the variationswithin video scenes can significantly influence the burstinesslevel of the stream [20], [21]. The burstiness of a VBRstream traffic increases the complexity of network resourcesmanagement to ensure QoS support for continuous streamplayback. Although, the reference design of the HCCAscheduler is simple and efficient in supporting constantapplication profile, yet it is not adequate since it cannotaddress the fast-changing imposed by the VBR bursty traffic,which hinders the performance of HCCA by causing packetsto wait for a longer time in their transmission queues.In case of downlink traffic, from QAP to QSTAs, the QAPis aware about its data queues and shall use its highest priorityto seize the channel if it remains idle for a duration of PIFSwithout undergoing back-off procedure. However, due to thefact that QAP suffers from the lack of information aboutthe uplink transmission queue status, an adaptive schemeis required to allow the scheduler to adjust its behaviorbased on the current application characteristics. Generally,adaptation to the application can be categorized according VOLUME 0, 2018 . A. Al-Maqri et al. : Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e to its variability level-based in the three well-known typesmentioned in III-A.In MAC layer, the uplink traffic profile can be determinedusing different ways, such as estimating the data bufferof the flow, predicting the packet generation time and/ortraffic load at a specific time, or obtaining actual informationthrough cross-layer architecture design. By having the trafficprofile, the HCF can adjust one or more of its functionssuch as polling [22], SI assignment [23], TXOP allocationmechanisms [24] which allows it to instantaneously adapt toQoS requirement of the flow.The QoS of VBR video transmission is ungoverned dueto the fact that those packets are queued for a durationequivalent to SI until already-queued packets in the buffer aredelivered. Recall that during each SI, the reference HCCAscheduler allocates a fixed TXOP to each QSTA based onits mean rate requirements regardless the real VBR trafficchanges. There are three QoS challenges relevant to ClassI, II and III of VBR traffics.
1) QoS Challenges of Class I video flows
HCCA scheduler fails to accommodate to variability ClassI traffic which, in turn, leaves the wireless bandwidth inunderutilization status. Assume, without loss of generality,that an identical TXOP duration is allocated for every QSTA,consequently, each QSTA will waste the same amount ofunused TXOP ( T u ). Thus, Equation (5) can be rewritten asfollows: T XOP n +1 SI + T XOP − T u SI ≤ T − T CP T (6)According to Reference [25], using different SIs fordifferent streams will improve the bandwidth utilization upto 50%. In other words, the T u in Equation (5) will be equalto T XOP . Therefore, the Equation (5) can be again rewrittenas follows: T XOP n +1 SI + T XOP × SI ≤ T − T CP T , (7)which means that the number of admitted flows can bemaximized to double the number of admitted flows whendifferent SIs are used.
2) QoS Challenges of Class II video flows
In Class II, when
QST A i , at any SI, exploits only portionof its allocated T XOP i at the traffic setup time, namely T ieff , leaving an unspent amount of T iu . Thus, the followingrelation can be held [26]: N (cid:88) i =1 T (cid:48) i = T eff + T eff + · · · + T Neff = T D − T u + T D − T u + · · · + T D N − T Nu = N (cid:88) i =1 T D i − N (cid:88) i =1 T iu (8) where T iu ≥ , (cid:80) Ni =1 T D i and (cid:80) Ni =1 T (cid:48) i is the total TXOPscheduled in any SI used in HCCA and ATXOP, respectively.It is worth noting that T D i is the TXOP duration of the QST A i including the poll overhead. Thus, the delay of QST A i in an SI is computed as follows: D iSI = i − (cid:88) j =1 ( T D i − T iu ) + T iL + T poll + 2 × SIF S (9)Altogether, the real QoS challenge is to minimize packetdelay by minimizing the surplus amount, namely T iu .
3) QoS Challenges of Class III video flows
In video streams like H.263, the deviation comprises not onlypacket size but also shows up to high variation in generationinterval which makes the matter much worse. In any SI,scheduling a QSTA based on its TSPEC likely imposesallocating surplus of TXOP duration which leads to wastingof the resources. This waste of resources due to the variationsin data rate influences the efficiency of the scheduler thatdoes not implement any recovery policy. Besides, due tothe variation in the packet generation interval, perhaps thereare some QSTAs that are not ready to transmit which willbe considered as over-polling state. This waste of resources,due to the variations in data rate, influences the efficiency ofthe scheduler that does not implement any recovery policy.Overall, it hinders the meet of delay bounds requirements,which leads to a degradation in QoS provisioning.Consider the example illustrated in Figure 5 wherefour QSTAs are polled for transmission in both CFP andControlled Access Phase (CAP). In this example,
T XOP , T XOP , T XOP and T XOP to QST A , QST A , QST A and QST A , respectively. The wasted TXOPand over-polling issues experienced using reference HCCA,inspired from the example, are as illustrated in Figure 5. • Over-polling of QSTAs
As illustrated in this example,due to the lack of awareness about the change inthe traffic profile, some QSTAs may receive unwantedpoll messages as their transmission queues are empty.
QST A and QST A in CP, and QST A in CFP willrespond with a null-frame causing unwanted delay to allQSTAs that may come after them in the same SI. • Wasted TXOP duration
Since some
QST As , such as
QST A , experience a high instant drops-down in datarate, only a short amount of the given TXOP durationis utilized. In this case, the channel might remain idlefor a period of time greater than the Short Inter FrameSpace (SIFS) and the control of the medium conveyedto Access Point (AP) to poll the next station in the list.Even though the effect of wasted TXOP duration in thepacket delay is not as high as that caused by over-pollingcase, however, it is considerably can go high as thenumber of stations in the network increases. VOLUME 0, 2018
5. A. Al-Maqri et al. : Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e
FIGURE 5.
Wasting TXOP and poll issue with VBR traffic transmission
B. ADAPTATION TO VARYING NETWORK CONDITIONS
Due to the phenomena of path loss, multipath fading,shadowing, and interference, wireless networks likely sufferfrom Signal-to-Interference Plus Noise Ratio (SINR) [23].The fluctuation of the underlying channel capacity willhinder the QoS provisioning for time-sensitive applications.Consequently, two possible ways to be applied on the QoSalgorithms in order to encounter this challenge and meetthe required QoS needs. The first one is by computing thetransmission time for the packets based on the minimumphysical bit rate announced. By doing so, the QoS isguaranteed, however, this technique gives rise to degradationin bandwidth efficiency as the bandwidth might get higheranytime while only the minimum link rate is considered. Thesecond one is to encourage the QoS algorithm to take intoaccount the link adaptation mechanism of WLANs over thetime.Although, piggybacking feature of HCCA is basicallydesigned to improve the channel capacity, it may inverselybehave when a station experience successive retransmissionor channel noise. This issue has been referred to in [27] as"the piggyback problem as the low physical transmissionrate". If any QSTA was transmitting at a low physicalrate due to channel error, QAP will accordingly decreasethe transmission rate of the piggybacked Contention Free(CF-Poll) frame. This, in turn, will result in channelefficiency degradation and will increase the TSs’ frame delayof other stations involving in the Network Allocation Vector(NAV) process.In case of VBR traffic transmission over WLANs andapart from the issues and challenges of HCCA reported in[18], [28], the major issue of the reference scheduler isthe unawareness about the inherent wireless time-varyingchannel condition[29]. Keeping aware about the channelstatus has a major impact on the scheduler performance asit can potentially degrade the service differentiation process,even though, HCCA has been observed to perform wellin heavy loaded network [30], [31], especially with theemergence of several physical layer technologies such asAdaptive Modulation and Coding (AMC) schemes.
C. BANDWIDTH UTILIZATION
In HCCA, after receiving an ADDTS-request from thestation, the scheduler needs first to calculate the requiredTXOP duration taking into account its TSPEC parameters.Thereafter, the used admission control mechanism will checkthe ability to accept the new TS. If the new TS is accepted, theSI will be computed as the minimum among all delay boundsof admitted streams, which is enough to meet the most urgentdelay requirement to guarantee the required QoS service forthe admitted streams. Finally, the round robin approach isused to allocate TXOPs to the involved station. Even thoughthe use of this design is very simple and straightforward,it still suffers from some challenging issues related to theefficient use of the bandwidth. Indeed, the use of round robinapproach in HCCA scheduler to serve all TSs in one SI mightlead to over-allocating the bandwidth, which in turn leads tounder-utilizing the channel bandwidth. Moreover, the wasteof the wireless bandwidth may reach up to more than 50% insome cases [32].In fact, based on the minimum physical rate and thecharacteristics of the incoming TSs, the ACU decides thenumber of admitted TSs to which the wireless resource willbe allocated. This approach leads to allocating a constantamount of resource to every TS using the mean of singlephysical transmission rate, which is not compatible withthe condition of current wireless bandwidth, especially VBRtraffic. In other words, the ACU should consider both thephysical layer and the service specific QoS parameters inorder to be able to achieve effective bandwidth utilization[33].With noticeably VBR flows, one of two scenarios likelyoccurs at some specific SIs. In the first scenario, the datarate becomes lower than the average value determined in theTSPEC, thus, the allocated TXOP will not be completelyconsumed which is considered as a wasting of resources.As for the second scenario, the data rate becomes greaterthan the average value determined in the TSPEC, thus, theassigned TXOP will not be enough to transmit the relevantdata which increases the end-to-end delay of the flow. Thepossible solutions to solve these two problems as explainedin these references [34], [35] are: (1) By increasing theTXOP duration to the average TXOP of traffic for the first VOLUME 0, 2018 . A. Al-Maqri et al. : Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e case, knowing that it will reduce the bandwidth utilization,especially if the data rate is dropped down. (2) By applyingthe bandwidth reclaiming approach [36], [37], [38], [39].
D. NETWORK RESOURCES MANAGEMENT
Indeed the HCF of IEEE802.11e protocol is targeted to theprovisioning of QoS throughout the service differentiation,yet the proper network resource management, such ascoordinating between distributed (CP) and controlled (CAP)periods and link layer resources still in request [40]. Inaddition, a feasible ACU scheme is also required, whichin such way can ensure that the QoS requirements aresatisfied. The HCCA scheduler operates based on the staticconfiguration of its traffic TSPEC parameters where theyare constantly served for their lifetime to enforce resourcesharing with ensuring that the desired QoS constraints aremet. To this aim, a good resource utilization is often leftto the heuristic network administrator know-how. However,this constant resource sharing policy might highly cause ascarce bandwidth utilization since it cannot adapt to thetransformation of the traffic profile and the lifetime due todynamic VBR traffic evolution.As a resolution to this issue, a bandwidth sharing strategyis suggested to rely on a criteria which is driven by theperformance [41], in which a common performance metricis recommended to be defined to differentiate between thetraffic streams based on their performance requirements.
IV. CLASSIFICATION OF QOS SUPPORT FORMULTIMEDIA TRAFFIC APPROACHES IN IEEE802.11EWLAN
In general, the enhancement approaches in IEEE802.11eprotocol can be classified based on the access mediumcontrol fashion into distributed control and centralizedcontrol enhancements. In IEEE802.11e, EDCA operatesbased on the distributed access control while the HCCArepresents the centralized access control. In [42], many QoSenhancements for 802.11 WLAN have been proposed andclassified along with their advantages and disadvantages.Another survey in [43] has focused on the QoS provisioningin both EDCA and HCCA over IEEE802.11e networks. TheHCCA enhancement approaches can be themselves classifiedinto different categories according to several aspects such asthe functional, structural, environmental and location aspects.In [44] and [45], the authors presented a survey of variousadmission control in IEEE802.11e and they classify schemesbased on several aspects such as Measurement-Based,Model-Based and Hybrid schemes. In [46], the delay-EDDbased scheduler has been compared to the feedback controlbased scheduler in order to provide a better comprehensionabout the so-called packet scheduling in 802.11 WLANs.Below is a short description of the different possible waysof IEEE802.11e approaches classification. • Traffic flow direction:
In infrastructure mode ofIEEE802.11 WLANs, the traffic directions would beeither downlink and uplink. "Downlink" refers to a traffic flow transmitted from AP to a mobile device,while "uplink" refers to a flow with a reverse direction.IEEE802.11e enhancements can be tailored to enhancethe performance for either downlink or uplink traffic orin some cases for both directions. • Targeted environment:
Although IEEE802.11e MACwas originally designed for wireless infrastructurenetworks and widely used in WLANs, there havebeen some enhancements for adapting IEEE802.11e towork with other networks such as the improvementof polling and scheduling scheme over IEEE802.11a/e[47], IEEE802.11p networks [48], Ad-hoc WirelessNetworks in [49] or in Integrated model of IEEE802.11eand IEEE802.16 [50], [51]. • Delay-EDD based and feedback control based:
The pre-knowledge of packets arrival time is onlypossible for the downlink. While in the uplink, neitherthe delays of the head of line packets nor the quotaof bandwidth needed by each flow are possible tobe known by the access point. For this reason,IEEE802.11e schedulers have been categorized intothe earliest due date and the feedback control class.A thorough comparison between these types has beenpresented in [46]. • Layered vs. cross layer:
IEEE802.11e enhancementscan be introduced in two structures, cross-layer andlayered approaches. The cross-layer approaches rely oninteractions between two layers of the OSI architecture.These approaches were motivated by the fact thatproviding lower or higher layer information to MAClayer to perform better. The layered approaches relyon adapting OSI layers independently of the otherlayers. Cross-layer is a promising direction to improvethe overall performance of WLAN since it takes intoaccount the interactions among layers [52]. Thus,several enhancements [53], [54], [55] prefer to usecross-layer design for obtaining accurate informationfor scheduling purposes. • Technique or mechanism used:
The HCCAscheduling approaches can be classified based onthe techniques and/or mechanisms used in the design.In the literature, a diverse techniques were developedfor HCCA scheduling to boost its performance formultimedia transmission over error-prone WLANssuch as estimation based approaches [23], [56],predicting traffic profile [57], [58]. Moreover, someof these approaches modified one or more of HCCAmechanisms such as TXOP assignments [59], [60],[61], [62], polling mechanism [63], [64], [65] or ACU[54], [66], [67], [33]. • Analysis method used:
The approach might beanalyzed and/or evaluated using one of three methods,namely analytical model [68], [69], [70], [71];simulation experiment [72], [73], [74] and test bed [75].It is worth noting that the analytical model usually isdone to capture the characteristics and the shortcoming
VOLUME 0, 2018
7. A. Al-Maqri et al. : Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e of the approach and prior to the proposal of a solution.Although, in simulation and test bed methods usedfor evaluating the proposed scheme, they might becarried out to provide a preliminary study to investigatea particular issue in the existing scheme for possibleremedy.
V. QOS ENHANCEMENTS IN IEEE802.11ECONTROLLED ACCESS MODE
This section presents some of the leading approachesproposed in the literature to improve the QoS provisioningfor multimedia traffic. More emphasis has been put on thetransmission of VBR video streams in IEEE802.11e WLAN.The approaches are classified into six sets based on thestrategy used to improve HCCA performance. However,some approaches can be matched to several types ofstrategies, but are only classified to their main strategy.Moreover, in the layered approaches the focus was onlyon the enhancements of HCCA at the MAC layer. Therepresentative approaches are defined and their mechanismsare described along with a discussion concerning theirstrengths and weaknesses in improving QoS performance inIEEE802.11e WLAN. In addition, a comparison of the maincharacteristics of various HCCA approaches is provided foreach category. Besides some mathematical models that studyand provide insights to improve the HCF functions have beenpresented which can provide a promising avenue for furtherresearch and investigation.
A. HCCA POLLING ENHANCEMENTS
The polling mechanism of the legacy HCCA is responsiblefor the scheduling and the allocation of TS based on theirfixed reservations. Thus, the efficiency of this mechanismhighly depends on the accuracy of the flow specificationdeclared to the HC. Yet, as the flow profile of VBR mighthighly vary over the time, the allocation based on fixedreservations will cause degradation in quality of multimediaflows even when the channel resource is surplus. Moreparticularly, several issues may affect the efficiency ofthe HCCA such as the inefficient Round-Robin schedulingalgorithm, the overhead induced by the poll frames, and thelack of coordination between the APs of the neighboringBSS. The representative approaches that address these issuesand even more are summarized as follows.
CP-Multipoll is a robust multipolling mechanism aimsto increase the channel utilization and to minimize thecorresponding implementation overhead, which can berobust in error-prone environments like WLANs [76].Moreover, the proposed scheme provides a polling scheduleto ensure the bounded delay requirements of real-time trafficand it also provides an admission control mechanism. Themain aim of this scheme is to design an efficient pollingmechanism, due to its high impact on the performance ofHCCA, which is able to serve both CBR and VBR real-timetraffic. Unlike SinglePoll schemes where every STA receivesone poll frame when polled, CP-Multipoll aggregates many polls in a single multipolling frame incorporating theDCF into the polling mechanism. The frame format ofCP-Multipoll scheme is as shown in Fig. 6.
FIGURE 6.
CP-Multipoll frame format
Basically, CP-Multipoll conveys the polling order intothe contending order. This can be achieved by assigningdifferent back-off values to the streams in the polling listwith accordance to their ascending order in the pollinglist and allow the back-off to execute as soon as theyreceive the CP-Multipolling frame. Besides minimizing thepolling overhead by transmitting one polling message forall QSTAs in the polling list instead of sending polls asmany STAs, the proposed scheme has other advantagesover other multipoll schemes. The bursty traffic is bettersupported since the STA holds the channel only for a periodneeded to transmit its local buffered data. Moreover, in DCFaccess mode, if the STA does not use the poll frame dueto empty data buffer, the other STAs in this polling groupwill immediately detect that the channel is idle and it willadvance the starting of channel contention. However, theproposed mechanism is prone to hidden terminal problemssince each STA will decrement back-off counter when itsenses that the channel is idle. Thus, if hidden terminalsexist in the network, different STAs will complete theirback-off simultaneously and collision will happen. Due tothe inherent hidden node issue of infrastructure wirelessnetworks, CP-Multipoll cannot guarantee that all STAs inthe BSS can sense the transmission of other STAs. In thiscase, the station will transmit its data immediately upon theexpiration of its back-off timer leading to a collision.
CF-Poll piggyback scheme is presented by Lee and Kim[77] to optimize the usage rule of the CF-Poll piggybackscheme as defined in the IEEE802.11 standard [1] accordingto the TS load and the minimum physical transmission rate ofa QSTA which suffer the deep channel fading. Consider thecase of piggybacking, the CF-Poll in the QoS-ACK framefrom QAP to
QST A , illustrated in Fig. 7, must be listenedby all QSTA in the BSS. If any of the QSTAs experience lowphysical rate, which implies that QST A requires more timeto receive the frame, the delay for all other QSTAs will beincreased and the channel efficiency will be decreased.Motivated from the aforementioned issue, the proposedwork provides a guideline for the optimal usage ofthe CF-Poll piggyback scheme in IEEE802.11e andIEEE802.11n protocols. Simulation-based results revealthat the frame transmission delay is majorly affected by theminimum physical rate when CF-Poll is piggybacked in theQoS data frame while it is slightly influenced by the trafficload. The results show an inverse relationship between theCF-Poll piggyback scheme and the traffic load. Despite the VOLUME 0, 2018 . A. Al-Maqri et al. : Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e
FIGURE 7.
CF-Poll piggyback issue with an example of piggybacking CF-Pollon data frame
FIGURE 8.
Collision due to polling the STAs in the overlapping area presented analysis and guidelines, the recommendationsreckon on a number of assumptions that are: the trafficis Constant Bit Rate (CBR) and each QSTA has only oneTS calculated based on the Equation (3) which cannotbe suitable for supporting the transmission of multimediaapplications with variable profile.
Deterministic Back-off (DEB) method for HCCA isan enhancement of HCCA which performs virtual pollingthrough sensing the carrier of the wireless channel [78]. Thistechnique highlighted the issue of the collision incurred dueto polling the nodes in the overlapping area of two adjacentBSSs at the same time. This actually occurs due to the lack ofcoordination in HCCA between the adjacent APs. Considerthe nodes 5, 6, 7, 8 in the overlapping area illustrated inFig. 8. Since AP A cannot hear AP B, therefore the collisionoccurs between the nodes in the overlapped area. DEB uses asimilar idea of sensing the carrier of EDCA since it manifestshigh robustness and flexibility controlling the medium at theoverlapping BSSs. A virtual polling has been achieved in adistributed manner. The DEB arranges the back-off timer ofstation to guarantee that the polled stations will have differentback-off. When the back-off timer expires, the station canbe polled without colliding with others. However, DEB isonly functioning in CFP whereas HCCA is supposed to workin both CFP and CP, for this reason one of the significantmerits of HCCA will be untapped. Moreover, there is no clearconsideration of the readiness of the station, STA with nodata ready to send will be given a TXOP which, in turn, bewasted.
Non-Polling based HCCA (NPHCCA) is presented in[31] to provide an enhancement over HCCA mechanism.Since the VBR traffic exhibits variability in packet generation
FIGURE 9.
NPHCCA mechanism time, the station will not always have pending data totransmit, thus, it will waste time for the AP to send pollingmessages to the stations that have no data to transmit. For thisreason, the proposed solution modifies the HCCA scheme insuch way it allows stations that have pending frames to reporttheir readiness status to AP through exchanging messages.Then, the AP schedules the only ready stations in appropriatetransmission sequence.The mechanism of the NPHCCA is carried out throughouta sequence of messages exchanging. First, a station withdata will send a transmission frame request to the APin order to update it about its transmission queue status,including information such as required Priority, Queue status,etc. A station only sends this frame after it receives thebeacon message from the AP and senses whether themedium is idle for SIFS. Accordingly, the AP maintainsthis information in its scheduling table. Finally, the APdetermines a transmission sequence and notifies stationsto transmit data according to this transmission sequencebroadcast in the beacon messages. Fig. 9 demonstratesThe components of the NPHCCA. Although, NPHCCA hasshown improvement in the transmission delay when thenetwork is light-loaded, the performance was similar to thatof HCCA when the network is heavy-loaded. Besides, themessaging exchange of the beacon and transmission requestframes added extra overhead to the network, especially whenthe number of the nodes increases.
F-Poll
In Feasible Polling Scheme (F-Poll) [79], theapplication layer gives the accurate arrival-time of theupcoming data frame over the uplink connection to the MAClayer, where this approach is known as a cross-layeringapproach. F-Poll is suitable for both type II and III ofvideo types categorized in Subsection III-A, where the exactinformation of the next inter-arrival time is sent to the QAPin order to enhance the scheduling of the TSs. In order toavoid polling a station that have no ready data to transmit,a decision is made of whether to poll the relevant stationin the upcoming SI or not directly after receiving a dataframe. As a result, the packet access delay is minimizedand a great amount of unused TXOP duration is conservedwhich efficiently enhances the channel utilization. Fig. 10elaborates the F-Poll Mechanism.
AMTXOP [26] like D-TXOP [80], the AdaptiveMultipolling TXOP Scheme (AMTXOP) calculates theTXOP for a certain data stream based on the actual framesize. Since the polling messages can increase the overheardamong all QSTAs, the BSS broadcasts one multi-polling
VOLUME 0, 2018
9. A. Al-Maqri et al. : Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e
FIGURE 10.
F-Poll scheme mechanism message to the QSTAs in a single SI instead of sendingone polling message for each. This approach minimizes thepolling overhead and also reduces the packet delay whichsignificantly improves the bandwidth utilization. Due to thisintegration, the AMTXOP outperforms both HCCA and itsancestor, D-TXOP, in terms of channel utilization and packetdelay.
B. TXOP ALLOCATION ENHANCEMENTS
Usually, if a QSTA’s buffer queue is empty during its TXOPbecause of a non-uniform data flow from the upper layer,the media will be unutilized for the whole TXOP of thestation. However, according to the 802.11e standard, theQSTA should send a QoS-NULL frame to the QAP toenforce it to start polling other sessions immediately [71].On the other hand, if the allocated TXOP is not enoughto send the backlogged packets, these data will be servedin the next SI causing more delay and might impair thedesignated QoS requirements [38]. Several techniques havebeen presented in the literature to address the limitation of theTXOP assignment mechanism of IEEE802.11e, we overviewhere some representative approaches.
Scheduling Based on Estimated TransmissionTimes-Earliest Due Date (SETT-EDD) [23] has proposeda novel scheduling technique for the so-called IEEE802.11eHCF. A simple mechanism similar to the CAP timer hasbeen employed to limit the polling-based transmissionin HCF which so-called TXOP timer. This TXOP timerincreases at a constant rate equal to the proportion of thatTXOP duration to the minimum service interval (
T D/mSI ),which reflects the fraction of time consumed by the station inpolled TXOPs. The maximum value of this timer is equal tothe maximum TXOP duration (
M T D ). The consumed timeby a station in a polled TXOP is subtracted from the TXOPtimer by the end of the TXOP. Thus, the station can bepolled only if the TXOP timer value is greater than or equalto the minimum TXOP duration ( mT D ), which guaranteesthe transmission of at least one data frame at the minimumPHY rate.Since the TXOP is allocated in SETT-EDD based onearliest deadlines, the transmission delay and data loss havebeen reduced. That is why SETT-EDD shows flexibility andconsidered a representative dynamic scheduler, as well as it
FIGURE 11.
The ARROW Mechanism provides compatibility to the link adaptation implemented inthe commercial WLANs. However, it still lacks an efficienttechnique to be able to calculate the accurate required TXOPfor every QSTA transmission instead of estimating TXOPbased on the average data rate of each TS and the packet timeinterval between two consecutive transmissions.
Adaptive Resource Reservation Over WLANs(ARROW) is another algorithm where the TXOP assignmentis calculated dynamically based on the queued data sizeof the QSTAs [81], [82]. In ARROW, the SETT-EDD[23] has been extended, where the available bandwidthis allocated based on the existing amount of data whichis ready for transmission in every STA. In contrast toSETT-EDD, which allocates the channel bandwidth based onthe expected arriving data in every STA. In this mechanism,QSTA advertises the size of the total queued packetswaiting for transmission with every poll. This informationis piggybacked with the data frame prior the sending backto QAP. So, the next TXOP allocation for any particularstream will be calculated based on the advertised queuesize. In this algorithm, the channel is allocated based onthe exact transmission requirements for each QSTA, whichis expressed by the Queue Size (QS) field indicated duringthe previous TXOP. By doing so, the TXOP is assignedto meet the transmission requirement at the time when theprevious TXOP assignment is made and consequently thedata buffered in the QSTA is taken into account at any SIleading to efficient adaptation of bandwidth allocation toactual requirements. Specifically, as illustrated in Fig. 11,data arrive during [ t i ( x ) , t i ( x + 1)] can only be transmittedafter the elapsing of t i ( x + 2) , which results in a delay ofpackets for at least one SI. Enhanced Earliest Due Date (EDD) QoS scheduler: presented by [83] and it is an EDD-based algorithm mainlyaims at addressing the above-mentioned weakness of theARROW scheduler. Similar to ARROW scheduler, theEnhanced EDD also uses the queue length information likeARROW. However, the Enhanced EDD estimates the numberof arriving packets immediately after the end of the previoustransmission, as shown in Fig. 12. Thereafter, it calculatesjust the enough TXOP to clear up the buffer queue by theend of current transmission. To reduce the average delay,when the buffer is not empty after the current transmissioncompletes, the next SI begins earlier, which can be achieved VOLUME 0, 2018 . A. Al-Maqri et al. : Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e
FIGURE 12.
TXOP assignment in enhanced EDD by changing the value of mSI and Maximum Service Interval(MSI). The TXOP allocation in Enhanced EDD is calculatedfor each station
ST A i as the summation TXOP calculatedexactly as in ARROW and a duration enough to transmit thepackets generated during the current SI as below T XOP i = T XOP iavg + T Dr i (10)where T XOP avg is calculated as follows and N curSI is theexpected number of packets generated from time t pre until t (cid:48) . Dynamic TXOP HCCA Dynamic TXOP HCCA(DTH) involves a bandwidth reclaiming mechanism intoa centralized HCCA scheduler in order to improve thetransmission capacity and to provide additional resourcesto VBR TSs [34]. The main concept of DTH is to preventwasting the underutilized portion of transmission time inorder to allocate it to the next polled station that needs longertransmission period. This approach relies on the unspentamount of the TXOP from the previous poll time of a
QST A i as follows T XOP i = (cid:26) T XOP AC ( i ) if T spare ≡ t est ( i )+ T spare if T spare > (11)If there is no surplus TXOP duration from previous polltime, which implies that the station exhausts the whole TXOPduration. The next TXOP duration will be the same as the onecalculated in Equation (4). Otherwise, it will be calculatedas the summation of the unused TXOP duration and theestimated transmission time, computed through the SimpleMoving Average (SMA) of the effectively utilized durationin the previous polling intervals. Simulation results showthat this approach can improve the performance, especiallyin terms of transmission queue size, data loss and delay,and the approach can absorb and follow the variation ofVBR. Additionally, another analytical study confirms that theDTH approach has no effect on the policy of the centralizedscheduler. However, the estimation of transmission timeusing Moving Average needs more investigation as the VBRtraffic tends to high variability during the time, thus it mightbe not efficient to find the best setting of the mobile samplingwindows. The Dynamic TXOP (D-TXOP) scheduling algorithm [80] analyzes the video of the prerecorded streams before thecall setup, which has been previously highlighted in [84].The D-TXOP is suitable for transmitting type (I) of VBRvideo source categorized in Subsection III-A, which showsvariability in packet size. Indeed, this approach assigns theTXOP for a stream based on the real frame size rather than
FIGURE 13.
Dynamic TXOP assignment scheduling algorithm.
FIGURE 14.
TSPEC element format the estimated average of frame size. It uses the unused QSfield of IEEE802.11e MAC header to send the actual size ofthe upcoming frame to the HC. Thus, the wasted TXOPs havebeen minimized by this approach, which reflects lower delaycompared to the previous solutions. Moreover, the EDCAbenefits from the surplus TXOP duration from unused TXOPof the preceding STAs as illustrated in Fig. 13.
C. HCCA ADMISSION CONTROL ENHANCEMENTS
The main purpose of HCF admission control is to administerpolicy or regulate the available bandwidth resources whichis used by the HC. The admission control is used tolimit the amount of traffic admitted under a certain servicecategory in order to guarantee the highest possible QoS level,while maximizing the utilization of the medium resources.Fig. 14 depicts a common frame format for carrying TSPECparameters. Since the admission control relies on a fixedTSPEC element, it cannot efficiently cope with the highvariability of VBR streams. To solve this problem, numerousenhancements and optimizations have been proposed totackle this deficiency in the legacy ACU mechanism.
Rate-Variance-envelop-based Admission Control(RVAC) mechanism uses the Dual Token Bucket (DTB)shaper to guarantee the desired QoS specification [85]. Theauthors of these two references [86], [87] have derived thedelay probability based on the aggregate traffic statisticsrather than considering each flow individually to accepta new flow for admission [88], [89]. The effective TXOPduration of a recently arrived VBR stream can be inverselyderived based on a given packet loss rate as in Equation (5).Indeed, the RVAC takes the multiplexing gain of VBRtraffic into account unlike the guarantee-rate-based scheme.More specifically, if the arrival time of data streams extendsover a wide range, where the RVAC can fully utilize themultiplexing gain among the VBR streams, the performance
VOLUME 0, 2018
11. A. Al-Maqri et al. : Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e gain can be noticeable. Additionally, the RVAC considersboth uplink and downlink traffic streams. Simulation resultshave shown that the admission capacity of the RVACapproach is more than the double of its equivalent in theGRAC approach. In addition, the RVAC scheme will notviolate the 0.1 second delay requirement as long as thestarting time of the streams are spread over a wide timerange of not less than 2 seconds. However, the performanceof the RVAC in the wireless channel errors environment isnot studied.
Equal-SP [90] has been designed of HCCA scheduling,in which the spacing of a particular stream is determinedas the period of time between two consecutively scheduledstreams. It has been called equal-SP scheduling becausea particular stream will always get an equal spacing forits scheduling slices in the schedule. Indeed, the equal-SPscheduling relies on the well-known Rate Monotonic (RM)algorithm to achieve the QoS requirements. Despite that theequal-SP approach is similar to the SETT-EDD in terms ofthe general scheduling concept, however, the former assignsequal spacing for each particular stream, which is proven toviolate the delay requirement in some cases.In the example as shown in Fig. 15, the scheduler assigned25 ms, 50 ms, and 150 ms time spicing for the flows 1, 2,and 3, respectively, which makes T = T = T = T =50 ms, and T =
150 ms. The equal-SP approachis easy to be implemented and it can guarantee the delayrequirements and efficiently utilize the bandwidth whilemaintaining the compatibility to the standard since it uses thesame TSPEC parameters. However, the equal-SP approachencounters the same issues faced by the standard; if a newlyadmitted stream has a smaller delay bound than the current T , the current T will be set to less than or equal to thenew delay bound. Therefore, the TXOP durations for thepreviously admitted flows are required to be recalculated withthe T i s. Additionally, the scheduler needs to reassign indexesto the admitted flows in order to maintain the condition of T ≤ T ≤ · · · ≤ T n . FIGURE 15.
An Example of the Equal-SP scheduling. The QoS is guaranteedby applying admission control
PHCCA , as described in Fig. 16, is a priority basedQoS and admission control used for queue managementmechanism [91]. In this approach, a mechanism forborrowing and returning bandwidth among queues hasbeen studied. The higher priority queue, called class, haspermission to borrow bandwidth from lower priority queueswith the awareness of starvation protection for each priorityqueue. PHCCA modifies the HCCA by classifying the traffic into 3 classes, which has not been divided by the standard.Class 1 is the highest priority class suitable for voice andconference traffic implementation. Class 2 is the secondhighest priority class suitable for broadcast video traffic.Class 3 is the lowest priority traffic suitable for FTP andHTTP traffic.
FIGURE 16.
PHCCA admission control mechanism
Experimental results reveal that the proposed PHCCA canaccept more requests for Class 1 (70% better) compared tothe regular HCCA. Meanwhile, the second highest priority(Class 2 in this case) is still served quite similar to the regularHCCA. Despite that the PHCCA significantly outperformsthe regular HCCA, it is still not able to guarantee the requiredQoS for every flow; since it assumes that flows of similartypes (e.g. VoIP) will have exact QoS requirements. Besides,the performance of NPHCCA could merely achieve theperformance level of Best-Effort for VoIP and CBR-videoapplications. Moreover, parameters and environment factorsfor bandwidth borrowing mechanism, such as distance fromthe access point or mobility, should be investigated.
AF-HCCA [92] enhances the experienced delay of videotraffics by utilizing the surplus bandwidth and mitigatesthe over-polling issue. It computes the TXOP for a trafficstream based on the knowledge about the actual upcomingframe size instead of assigning TXOP according to the meancharacteristics of the traffic which is unable to reflect theactual traffic. This scheduler exploits the queue size field ofIEEE802.11e MAC header to transfer this information to theHC.In AF-HCCA, the QSTAs will be prevented from receivingunnecessary large TXOP which produces a remarkableincrease in the packet delay. Furthermore, the surplus timeof the wireless channel conserved by reducing the number ofpoll frames throughout the feedback is another benefit of thisresearch.The integrated scheme of AF-HCCA shows superiorperformance compared to IEEE802.11e HCCA, EnhancedEDD [83] and F-Poll [79] schedulers in terms of delay andchannel utilization without affecting the system throughput.However, preserved TXOP time is not efficiently utilized toenhance the flow capacity.
Feedback-based Admission Control Unit (FACU) [93]aims at maximizing the utilization of the surplus bandwidth VOLUME 0, 2018 . A. Al-Maqri et al. : Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e which has never been tested in previous schemes. The FACUexploits piggybacked information containing the size of thesubsequent video frames to increase the number of admittedflows.The FACU introduces an enhancement to admissioncontrol mechanism of Adaptive-TXOP. Analytical resultsreveal the efficiency of FACU over the examined schemes.The results show that the conserved channel bandwidthof Adaptive-TXOP can be utilized to increase the numberof admitted QoS flows and enhance the overall QoSprovisioning in IEEE802.11e WLANs.
VI. HCCA SCHEDULING APPROACHES COMPARISON
Table 4 presents a summary and comparison for the HCCAenhancements in IEEE802.11e along with their targetedfeatures classified based on the place of the enhancement.The solution column briefly describes the used technique.The complexity of an approach can be high, medium orsimple estimated based on Likert-type rating scale. Thecomplexity here represents the volume of the operationsof that particular approach. The method that involves moreoperations is considered high-complex and vice versa. Themain targeted traffic of the enhancement is stated. Thetargeted flow direction, which is considered by the approach,is also presented.
VII. OPEN RESEARCH ISSUES
Although the existing approaches provide several possiblesolutions to alleviate the deficiency of scheduling for VBRmultimedia traffic in IEEE802.11e WLANs, many issueshave been thoroughly discussed in the literature reviewsection, which are potential research topics. This sectionhighlights the most important issues in order to determine thedirections for potential future research. One of the problemswith HCCA is the coexistence. Several mechanisms claim tobe able to coordinate different HCs that operate on the samefrequency channel. Since HCCA’s QoS guarantee depends onthe exclusive usage of the frequency channel, multiple HCCAcan hardly coexist. On the other hand, additional delay mayoccur by the polling STAs with scalable video that exhibitsconstant quality yet introduce high variation in the trafficprofile. From the cross-layer perspective and to the best ofour knowledge, there is no proactive scheme that providesa good solution to the adequate interaction between thefluctuation of the uplink VBR traffic profile at the applicationlayer and the flexible scheduling policy at MAC layer whichexhibit low-complexity design. In summary, some issues areneeded to be considered to provide optimal enhancement forthe transmission of VBR traffic in IEEE802.11e WLANs. Webelieve the following suggestions are desirable for designinga good HCCA scheme in IEEE802.11e wireless networks. • Providing efficient estimation of the bandwidth in orderto achieve high connection throughput. • Designing a scheme coupled with link adaptationmechanisms in order to provide efficient adaptation todynamic network behavior. N u m b e r o f P ub li ca ti on s Year PollingTXOP AllocationACUHybridAnalysis
FIGURE 17.
Number of publication of the investigated research areas • Exploiting the distributed feature of EDCA in designinga hybrid HCCA scheme in order to yield highintegration and interoperability without jeopardizing thesystem simplicity. • Enabling the fragmentation and the blockacknowledgment introduced in the standard [11]with HCCA scheme. • Achieving low algorithm complexity.
VIII. RESEARCH TREND ON QOS SUPPORT INIEEE802.11E
Many researches have been conducted in the Literature sincethe first advent of the HCCA protocol draft in IEEE802.11estandard [94]. These researches can be classified into fiveresearch areas as in Table 5 aims at demonstrating the trend ofthe research since the first presence of the HCF functions till2009 and from 2010 to present. The collection includes over89 journal and conference papers. These scientific documentshave been collected using IEEE Explorer Digital Library,Springer Link, ScienceDirect and Google Scholar. One cannotice that the polling and TXOP allocation mechanismshave greatly received the researchers’ attention since theevolution of the HCCA till now, while admission controlmechanisms have less interest. It is worth noting that thedesign of the hybrid EDCA-HCCA scheme has scarcelystudied. The HCCA performance and mathematical analysishave been fairly covered. Yet, only few efforts have focusedon designing a comprehensive analytical model for HCCAprotocol. The aggregated number of papers published in threeperiods, namely 2004 to 2007; 2008 to 2011 and 2012 topresent are depicted in Fig. 17. The figure shows that thepolling and TXOP mechanism have received a great amountof attention compared to ACU and hybrid scheme. On theother hand, recently there has been a few analysis of HCCAprotocol, in contrast to the period from 2004 up to 2011.
IX. FUTURE DIRECTIONS
Although all proposed schemes in their current statesimprove the transmission of prerecorded video, there stillsome issues need to be addressed and investigated. Below,
VOLUME 0, 2018
13. A. Al-Maqri et al. : Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e
TABLE 4.
Comparison of The main characteristics of the HCCA approaches
Strategy Approach Solution ComplexityLevel ∗ Targeted Flow Direction Main QoS challengeCP-Multipoll [76] Multipolling scheme High Voice/video Uplink/downlink Packet delayHCCA polling CF-Poll [77] Piggyback Medium Voice/video Uplink Flows capacitymechanisms DEB method for HCCA[78] Deterministic polling Simple CBR/VBRvideo Uplink Flows capacityNPHCCA [31] Non polling feedback-based Simple Voice/video Uplink Packet delayF-Poll [79] Feasible Polling Scheme Simple Video Uplink Packet delay and flowcapacitySETT-EDD [23] Token bucket and Earliest Due Datebased Medium Voice andvideo Uplink/downlink Packet delayAMTXOP [26] Adaptive Multipolling TXOPScheme Simple Video Uplink Packet delay, flowcapacityTXOP allocation ARROW [81] Feedback based Simple Voice andvideo Uplink/downlink Flow capacitymechanism Enhanced ED [83] Estimation and feedback based Medium Voice andvideo Uplink Packet delayDynamic TXOP HCCA[34] Bandwidth reclaiming mechanism High Voice andvideo Uplink Packet delay, flowcapacityD-TXOP [80] The Dynamic TXOP Schedulingalgorithm Simple VBR traffic Uplink Packet delayRVAC [85] Dual token bucket (DTB) shaper Medium VBR traffic Uplink/downlink Flows capacityHCCA admissioncontrol Equal-SP [90] Equal spacing scheduling Simple Voice andvideo Uplink Flows capacityPHCCA [91] Priority based Simple Voice, video Uplink/downlink Flows capacityAF-HCCA [92] Adaptive Feedback-based HCCA Simple Video Uplink Packet delayFACU [93] Feedback-based Admission ControlUnit Simple Video Uplink Packet delay, flowcapacity ∗ Note that the complexity level reflects the volume of the operations as explained in Section VI based on Likert-type rating scale
TABLE 5.
Researches in QoS provisioning of Multimedia traffic in IEEE802.11e
Area Published from 2004 to 2009 Published since 2010HCCA Polling [95], [65], [96], [63], [97], [77], [98], [99], [100] [78], [101], [31], [102], [103], [48], [104], [105], [79], [106]TXOP Allocation [56], [18], [66], [107], [81], [108], [109], [25], [110], [111], [112], [54], [29], [39], [55], [24], [113], [61], [34], [36][114], [38], [115], [62], [116] [59], [60], [26] [92]Admission Control [88], [117], [118], [66], [119], [120], [121], [85], [122], [123] [33], [124], [91], [125], [126], [60], [127], [93]Hybrid EDCA-HCCA [128], [129], [130], [131] [132], [73], [72]HCCA Analysis [133], [134], [135], [68], [136], [137], [138], [139], [38], [69], [140], [141], [142], [143], [144], [145], [70], [71][146], [147], [148] we highlight some future works that need to be carried outfor further enhancement to: • Enhance the HCCA to cope with more complicatedwireless scenarios, where the hidden node problemexists and QSTAs communicate using RTS/CTSmechanism with MAC level fragmentation andmulti-rate support enabled. • Study the scalable HCCA MAC for video over wirelessmesh networks that are also scalable to a wider rangeof MAC settings to support more robust time-boundedmedia applications. • Design a new admission control algorithm to utilizethe excess bandwidth and to manage the availableresources among the HCCA, HCF and EDCA in order tomaximize the number of served streams or applicationsin the network. • Examine the performance of HCCA with the presenceof collision occurred in the overlapping area whenpolling stations among two neighboring BSSssimultaneously.
X. CONCLUSION
IEEE802.11e is aimed at providing stringent QoS supportto multimedia applications such as video streaming. The controlled based function of IEEE802.11e standard, HCCAscheduler, consider a fixed TSPEC for scheduling thetraffic while in fact the VBR traffic tends to changetheir characteristics such as data rate and packet sizeover the time. The inability of the IEEE802.11e MACprotocol to accommodate to the high fluctuation of VBRvideo profile motivates many researches to be conducted.Several enhancements have been made to alleviate theseshortcomings. These enhancements tend to address particularissues or applications by improving, in most cases, one of theHCF functions. In general, designing a robust HCF solutionthat provides an integrated solution for all traffic classes isstill a challenging task for future research.
REFERENCES [1] IEEE 802.11 Standard, “IEEE Standard for Information Technology-Telecommunications and Information Exchange Between Systems- Localand Metropolitan Area Networks- Specific Requirements- Part 11:Wireless LAN Medium Access Control (MAC) and Physical Layer(PHY) Specifications,” ANSI/IEEE Std 802.11, 1999 Edition (R2003),pp. 1–513, 1999.[2] ——, “IEEE Standard for Information Technology- Telecommunicationsand Information Exchange Between Systems-Local and MetropolitanArea Networks-Specific Requirements-Part 11: Wireless LAN MediumAccess Control (MAC) and Physical Layer (PHY) Specifications,” IEEEStd 802.11-1997, pp. i–445, 1997. VOLUME 0, 2018 . A. Al-Maqri et al. : Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e [3] IEEE 802.11a Standard, “Supplement to IEEE Standard for InformationTechnology- Telecommunications and Information Exchange BetweenSystems-Local and Metropolitan Area Networks-Specific Requirements.Part 11: Wireless LAN Medium Access Control (MAC) and PhysicalLayer (PHY) Specifications: High-Speed Physical Layer in the 5 GHzBand,” IEEE Std 802.11a-1999, pp. 1–102, 1999.[4] IEEE 802.11b Standard, “Supplement to IEEE Standard for InformationTechnology- Telecommunications and Information ExchangeBetween Systems- Local and Metropolitan Area Networks- SpecificRequirements- Part 11: Wireless LAN Medium Access Control (MAC)and Physical Layer (PHY) Specifications: Higher-Speed Physical LayerExtension in the 2.4 GHz Band,” IEEE Std 802.11b-1999, pp. i–90,2000.[5] IEEE 802.11c Standard, “IEEE Standard for InformationTechnology-Telecommunications and Information ExchangeBetween Systems-Local and Metropolitan Area Network-CommonSpecifications,” IEEE Std 802.1t-2001, pp. i–40, 2001.[6] IEEE 802.11d Standard, “IEEE Standard for InformationTechnology-Telecommunications and Information Exchange BetweenSystems-Local and Metropolitan Area Networks-Specific Requirement.Part 11: Wireless LAN Medium Access Control (MAC) and PhysicalLayer (PHY) Specification. Amendment 3: Specifications for Operationin Additional Regulatory Domains,” IEEE Std 802.11d-2001, pp. i–26,2001.[7] IEEE 802.11e Standard, “IEEE Standard for Informationtechnology–Local and metropolitan area networks–Specificrequirements–Part 11: Wireless LAN Medium Access Control (MAC)and Physical Layer (PHY) Specifications-Amendment 8: MediumAccess Control (MAC) Quality of Service Enhancements,” IEEE Std802.11e-2005 (Amendment to IEEE Std 802.11, 1999 Edition (Reaff2003), pp. 1–212, 2005.[8] IEEE 802.11f Standard, “IEEE Trial-Use Recommended Practicefor Multi-Vendor Access Point Interoperability Via an Inter-AccessPoint Protocol Across Distribution Systems Supporting IEEE 802.11Operation,” IEEE Std 802.11F-2003, pp. 1–67, 2003.[9] IEEE 802.11g Standard, “IEEE Standard for Information Technology-Telecommunications and Information Exchange Between Systems-Local and Metropolitan Area Networks- Specific Requirements PartIi: Wireless LAN Medium Access Control (MAC) and Physical Layer(PHY) Specifications,” IEEE Std 802.11g-2003 (Amendment to IEEE Std802.11, 1999 Edn. (Reaff 2003) as amended by IEEE Stds 802.11a-1999,802.11b-1999, 802.11b-1999/Cor 1-2001, and 802.11d-2001), pp. i–67,2003.[10] IEEE 802.11n Standard, “IEEE Standard for Information technology–Local and metropolitan area networks– Specific requirements– Part 11:Wireless LAN Medium Access Control (MAC)and Physical Layer (PHY)Specifications Amendment 5: Enhancements for Higher Throughput,”IEEE Std 802.11n-2009 (Amendment to IEEE Std 802.11-2007 asamended by IEEE Std 802.11k-2008, IEEE Std 802.11r-2008, IEEE Std802.11y-2008, and IEEE Std 802.11w-2009), pp. 1–565, 2009.[11] IEEE 802.11e Standard, “Information technology–Telecommunicationsand information exchange between systems Local and metropolitanarea networks–Specific requirements Part 11: Wireless LAN MediumAccess Control (MAC) and Physical Layer (PHY) Specifications,”ISO/IEC/IEEE 8802-11:2012(E) (Revison of ISO/IEC/IEEE8802-11-2005 and Amendments), pp. 1–2798, 2012.[12] ITU-T, “ITU-T Recommendation E. 800 (08/94): Terms and definitionrelated to quality of service and network performance includingdependability,” 1994.[13] V. Jacobson, R. Frederick, S. Casner, and H. Schulzrinne, “RTP: Atransport protocol for real-time applications,” IETF RFC3550, 2003.[14] I. Busse, B. Deffner, and H. Schulzrinne, “Dynamic QoS control ofmultimedia applications based on RTP,” Computer Communications,vol. 19, no. 1, pp. 49–58, 1996.[15] ITU-T, “ITU-T Recommendation X. 200, (07/94): Informationtechnology–Open Systems Interconnection–Basic Reference Model: Thebasic model,” 1994.[16] J. Delsing, “Communication Technology in Mobile and PervasiveComputing,” Mobile and Pervasive Computing in Construction, pp.26–36, 2012.[17] M. M. Alani, Guide to OSI and TCP/IP Models. Cham: SpringerInternational Publishing, 2014, ch. OSI Model, pp. 5–17.[18] I. Inan, F. Keceli, and E. Ayanoglu, “An Adaptive Multimedia QoS Scheduler for 802.11e Wireless LANs,” in IEEE InternationalConference on Communications, vol. 11, 2006, pp. 5263–5270.[19] Jeng-Ji Huang, Yeh-Horng Chen, and Che-Yu Chang, “An MSI-BasedScheduler for IEEE 802.11e HCCA,” in IEEE 70th Vehicular TechnologyConference, 2009, pp. 1–5.[20] E. P. Rathgeb, “Policing of realistic VBR video traffic in an ATMnetwork,” International Journal of Digital and Analog CommunicationSystems, vol. 6, no. 4, pp. 213–226, 1993.[21] S. Bhattacharyya, S. Agrawal, A. Jeevani, and S. Sengupta, “Burstinessminimized rate control for high resolution H.264 video conferencing,” in2014 Twentieth National Conference on Communications (NCC), 2014,pp. 1–6.[22] N. Ramos, D. Panigrahi, and S. Dey, “Dynamic adaptation policiesto improve quality of service of real-time multimedia applications inIEEE 802.11e WLAN Networks,” Wireless Networks, vol. 13, no. 4, pp.511–535, 2007.[23] A. Grilo, M. Macedo, and M. Nunes, “A scheduling algorithm forQoS support in IEEE802.11 networks,” IEEE Wireless Communications,vol. 10, no. 3, pp. 36–43, 2003.[24] A. Jansang and A. Phonphoem, “Adjustable TXOP mechanism forsupporting video transmission in IEEE 802.11e HCCA,” EURASIPJournal on Wireless Communications and Networking, vol. 2011, no. 1,pp. 1–16, 2011.[25] Qinglin Zhao and D. Tsang, “Enhancing QoS Support in IEEE802.11e HCCA,” in IEEE Global Telecommunications Conference(GLOBECOM ’07), 2007, pp. 4909–4914.[26] M. A. Al-Maqri, M. Othman, B. M. Ali, and Z. M. Hanapi, “Adaptivemulti-polling scheduler for qos support of video transmission in ieee802.11 e wlans,” Telecommunication systems, vol. 61, no. 4, pp.773–791, 2016.[27] Hyun-Jin Lee, Jae-Hyun Kim, and Sunghyun Cho, “A Novel PiggybackSelection Scheme in IEEE 802.11e HCCA,” in IEEE InternationalConference on Communications, 2007, pp. 4529–4534.[28] S. Madhar Saheb, A. Bhattacharjee, P. Dharmasa, and R. Kar, “Enhancedhybrid coordination function controlled channel access-based adaptivescheduler for delay sensitive traffic in IEEE 802.11e networks,” IETNetworks, vol. 1, no. 4, pp. 281–288, 2012.[29] A. Arora, S.-G. Yoon, Y.-J. Choi, and S. Bahk, “Adaptive TXOPallocation based on channel conditions and traffic requirements in IEEE802.11e networks,” IEEE Transactions on Vehicular Technology, vol. 59,no. 3, pp. 1087–1099, 2010.[30] C. Cicconetti, L. Lenzini, E. Mingozzi, and G. Stea, “A softwarearchitecture for simulating IEEE 802.11e HCCA,” in IPS-MoMe05:Proceeding from the 3rd Workshop on Internet Performance, Simulation,Monitoring and Measurement, 2005, pp. 97–104.[31] Yeong-Sheng Chen, Yuan-Wei Lee, and Jong Hyuk Park, “EnhancedHCCA mechanism for multimedia traffics with QoS supportin IEEE802.11e networks,” Journal of Network and Computer Applications,vol. 34, no. 5, pp. 1566–1571, 2011.[32] Q. Zhao and D. H. Tsang, “Effective bandwidth utilization in IEEE802.11e,” in The Fourth International Conference on HeterogeneousNetworking for Quality, Reliability, Security and Robustness andWorkshops, 2007, pp. 1–7.[33] H.-J. Lee, K.-H. Lee, and J.-H. Kim, “A QoS provisioning mechanismsbased on effective bandwidth for the polling based WLAN system,”in Proceedings of the 4th International Conference on UniquitousInformation Management and Communication, 2010, pp. 1–10.[34] G. Cecchetti, A. L. Ruscelli, A. Mastropaolo, and G. Lipari, “DynamicTXOP HCCA Reclaiming Scheduler with Transmission Time Estimationfor IEEE 802.11e Real-Time Networks,” in Proceedings of the 15thACM international conference on Modeling, analysis and simulation ofwireless and mobile systems, 2012, pp. 239–246.[35] A. L. Ruscelli and G. Cecchetti, “A IEEE 802.11e HCCA Schedulerwith a Reclaiming Mechanism for Multimedia Applications,” Advancesin Multimedia, vol. 2014, pp. 1–22, 2014.[36] G. Cecchetti, A. Ruscelli, A. Mastropaolo, and G. Lipari, “ProvidingVariable TXOP for IEEE 802.11e HCCA Real-Time Networks,” inWireless Communications and Networking Conference (WCNC), 2012,pp. 1508–1513.[37] R. Lo Cigno, L. Palopoli, and A. Colombo, “Analysis of differentscheduling strategies in 802.11e networks with multi-class traffic,”in 32nd IEEE Conference on Local Computer Networks, 2007, pp.455–462.
VOLUME 0, 2018
15. A. Al-Maqri et al. : Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e [38] M. Rashid, E. Hossain, and V. Bhargava, “Controlled Channel AccessScheduling for Guaranteed QoS in 802.11e-Based WLANs,” IEEETransactions on Wireless Communications, vol. 7, no. 4, pp. 1287–1297,2008.[39] A. L. Ruscelli, G. Cecchetti, A. Mastropaolo, and G. Lipari, “A greedyreclaiming scheduler for IEEE 802.11e HCCA real-time networks,” inProceedings of the 14th ACM international conference on Modeling,analysis and simulation of wireless and mobile systems, 2011, pp.223–230.[40] N. Ramos, D. Panigrahi, and S. Dey, “Quality of service provisioning in802.11e networks: challenges, approaches, and future directions,” IEEENetwork, vol. 19, no. 4, pp. 14–20, 2005.[41] J. Navarro Ortiz, J. M. Lopez Soler, and G. Steay, “Quality of experiencebased resource sharing in IEEE 802.11e HCCA,” in European WirelessConference (EW), 2010, pp. 454–461.[42] Q. Ni, L. Romdhani, and T. Turletti, “A survey of QoS enhancementsfor IEEE 802.11 wireless LAN,” Wireless Communications and MobileComputing, vol. 4, no. 5, pp. 547–566, 2004.[43] H. Luo and M.-L. Shyu, “Quality of service provision in mobilemultimedia-a survey,” Human-centric computing and informationsciences, vol. 1, no. 1, pp. 1–15, 2011.[44] D. Gao, J. Cai, and K. N. Ngan, “Admission control in IEEE 802.11ewireless LANs,” Network, IEEE, vol. 19, no. 4, pp. 6–13, 2005.[45] Y. Liu and M. Meng, “Survey of admission control algorithms inIEEE 802.11e wireless LANs,” in International Conference on FutureComputer and Communication, 2009, pp. 230–233.[46] G. Piro, L. A. Grieco, G. Boggia, and P. Camarda, “QoS in wireless LAN:a comparison between feedback-based and earliest due-date approaches,”Computer Communications, vol. 35, no. 3, pp. 298–308, 2012.[47] M. Van der Schaar, Y. Andreopoulos, and Zhiping Hu, “Optimizedscalable video streaming over IEEE 802.11 a/e HCCA wireless networksunder delay constraints,” IEEE Transactions on Mobile Computing,vol. 5, no. 6, pp. 755–768, 2006.[48] B. ZHANG, M.-d. MA, C.-f. LIU, and Y.-t. SHU, “Improvement ofpolling and scheduling scheme for real-time transmission with HCCAof IEEE 802.11 p protocol,” The Journal of China Universities of Postsand Telecommunications, vol. 20, no. 3, pp. 60–66, 2013.[49] A. Fiandrotti, D. Gallucci, E. Masala, and E. Magli, “TrafficPrioritization of H.264/SVC Video over 802.11e Ad Hoc WirelessNetworks,” in Proceedings of 17th International Conference on ComputerCommunications and Networks, 2008, pp. 1–5.[50] V. Sattari Naeini and N. Movahhedinia, “Packet Scheduling andAdmission Control for QoS Provisioning in Integrated IEEE 802.11e andIEEE 802.16 Mesh Mode,” Arabian Journal for Science and Engineering,vol. 37, no. 6, pp. 1595–1611, 2012.[51] M. Lee and J. A. Copeland, “An adaptive polling algorithm withdifferentiation strategy in IEEE 802.16 multi-hop networks with IEEE802.11 e WLANs,” in IEEE 28th International Performance Computingand Communications Conference (IPCCC), 2009, pp. 354–359.[52] N. S. Shankar and M. van der Schaar, “Performance analysis of videotransmission over IEEE 802.11 a/e WLANs,” IEEE Transactions onVehicular Technology, vol. 56, no. 4, pp. 2346–2362, 2007.[53] C. Cicconetti, L. Lenzini, E. Mingozzi, and G. Stea, “An Efficient CrossLayer Scheduler for Multimedia Traffic in Wireless Local Area Networkswith IEEE 802.11e HCCA,” SIGMOBILE Mob. Comput. Commun.Rev., vol. 11, no. 3, pp. 31–46, 2007.[54] Z. A. B. M. Noh, T. Suzuki, and S. Tasaka, “Application-level QoSand QoE assessment of a cross-layer packet scheduling scheme foraudio-video transmission over error-prone IEEE 802.11e HCCA wirelessLANs,” IEICE transactions on communications, vol. 93, no. 6, pp.1384–1394, 2010.[55] H. Luo, “A Cross-Layer Design for Video Streaming Over 802.11eHCCA Wireless Network,” International Journal of Multimedia DataEngineering and Management (IJMDEM), vol. 2, no. 3, pp. 21–33, 2011.[56] Pierre Ansel, Qiang Ni, and Thierry Turletti, “An Efficient SchedulingScheme for IEEE 802.11e,” in In Proc. of WiOpt (Modeling andOptimization in Mobile, Ad Hoc and Wireless Networks, 2004, pp.391–403.[57] Sang-Jo Yoo, “Efficient traffic prediction scheme for real-time VBRMPEG video transmission over high-speed networks,” IEEE Transactionson Broadcasting, vol. 48, no. 1, pp. 10–18, 2002.[58] W.-K. Kuo and K.-W. Wu, “Traffic prediction and QoS transmissionof real-time live VBR videos in WLANs,” ACM Transactions on Multimedia Computing, Communications, and Applications, vol. 7,no. 4, pp. 1–36, 2011.[59] K. Ju and K. Chung, “Dynamic TXOP allocation for multimedia QoSproviding over wireless networks,” in 2013 International Conference onInformation Networking (ICOIN), 2013, pp. 397–401.[60] T.-H. Lee and Y.-W. Huang, “Quality of service guarantee for real-timeVBR traffic flows with different delay bound and loss probabilityrequirements in WLANs,” Journal of the Chinese Institute of Engineers,vol. 36, no. 4, pp. 471–487, 2013.[61] Z. A. M. Noh, M. N. M. Khambari, N. A. M. Ariff, and I. Roslan,“Retransmission-based additional TXOP allocation for audio-videotransmission by IEEE 802.11e HCCA,” in 7th International Conferenceon Information Technology in Asia (CITA 11), 2011, pp. 1–7.[62] Jeng-Ji Huang, Che-Yu Chang, and Huei-Wen Ferng, “FlexibleTXOP assignments for efficient QoS scheduling in IEEE 802.11eWLANs,” in 5th IFIP International Conference on Wireless and OpticalCommunications Networks, 2008, pp. 1–5.[63] Jenhui Chen and Chien-An Lin, “HMM: hybrid multipolling mechanismwith pre-allocation admission control for real-time transmissions inWLANs,” in 2004 IEEE 60th Vehicular Technology Conference, vol. 4,2004, pp. 3040–3044.[64] N. Ramos, D. Panigrahi, and S. Dey, “Dynamic adaptation policiesto improve quality of service of real-time multimedia applications inIEEE 802.11e WLAN Networks,” Wireless Networks, vol. 13, no. 4, pp.511–535, 2007.[65] Jungbo Son, Hosuk Choi, and Sin-Chong Park, “An effective pollingMAC scheme for IEEE 802.11e,” in IEEE International Symposium onCommunications and Information Technology, 2004, vol. 1, 2004, pp.296–301.[66] T. O. Kim, Y. Chang, Y.-T. Kim, and B. D. Choi, “An Admission Controland TXOP Duration of VBR Traffics in IEEE 802.11e HCCA withGuaranteed Delay and Loss,” in Proceedings of the 9th Asia-PacificInternational Conference on Network Operations and Management:Management of Convergence Networks and Services, 2006, pp. 162–169.[67] Jiang Zhu and A. Fapojuwo, “A new call admission control method forproviding desired throughput and delay performance in IEEE802.11ewireless LANs,” IEEE Transactions on Wireless Communications, vol. 6,no. 2, pp. 701–709, 2007.[68] S. Harsha, S. Anand, A. Kumar, and V. Sharma, “An Analytical Modelfor Capacity Evaluation of VoIP on HCCA and TCP File Transfersover EDCA in an IEEE 802.11e WLAN,” in Distributed Computing andNetworking, 2006, vol. 4308, pp. 245–256.[69] G. Cecchetti and A. L. Ruscelli, “Performance Evaluation of Real-timeSchedulers for HCCA Function in IEEE 802.11e Wireless Networks,”in Proceedings of the 4th ACM Symposium on QoS and Security forWireless and Mobile Networks, 2008, pp. 1–8.[70] A. Leonovich and H.-W. Ferng, “Modeling the IEEE 802.11e HCCAmode,” Wireless networks, vol. 19, no. 5, pp. 771–783, 2013.[71] A. Jansang and A. Phonphoem, “A Simple Analytical Model forExpected Frame Waiting Time Evaluation in IEEE 802.11e HCCAMode,” Wireless Personal Communications, vol. 69, no. 4, pp.1899–1924, 2013.[72] B. Ng, Y. Tan, and Y. Roger, “Improved utilization for jointHCCA–EDCA access in IEEE 802.11e WLANs,” Optimization Letters,vol. 7, no. 8, pp. 1711–1724, 2013.[73] A. L. Ruscelli, G. Cecchetti, A. Alifano, and G. Lipari, “Enhancement ofQoS support of HCCA schedulers using EDCA function in IEEE 802.11enetworks,” Ad Hoc Networks, vol. 10, no. 2, pp. 147–161, 2012.[74] B.-S. Kim, S. W. Kim, Y. Fang, and T. F. Wong, “Two-step MultipollingMAC Protocol for Wireless LANs,” IEEE Journal on Selected Areas inCommunications, vol. 23, no. 6, pp. 1276–1286, 2006.[75] A. C. H. Ng, D. Malone, and D. J. Leith, “Experimental Evaluation ofTCP Performance and Fairness in an 802.11e Test-bed,” in Proceedingsof the 2005 ACM SIGCOMM Workshop on Experimental Approaches toWireless Network Design and Analysis, 2005, pp. 17–22.[76] Shou-Chih Lo, Guanling Lee, and Wen-Tsuen Chen, “An efficientmultipolling mechanism for IEEE 802.11 wireless LANs,” IEEETransactions on Computers, vol. 52, no. 6, pp. 764–778, 2003.[77] H.-J. Lee and J.-H. Kim, “A optimal CF-poll piggyback scheme inIEEE 802.11e HCCA,” in 2006 8th International Conference AdvancedCommunication Technology, vol. 3, 2006, pp. 1954–1959.[78] Jeng-Ji Huang, Yeh-Horng Chen, and D. Shiung, “A four-way-pollingQoS scheduler for IEEE 802.11e HCCA,” in TENCON 2010-2010 IEEERegion 10 Conference, 2010, pp. 1986–1991. VOLUME 0, 2018 . A. Al-Maqri et al. : Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e [79] M. A. Al Maqri, M. Othman, B. M. Ali, and Z. M. Hanapi, “Feasiblehcca polling mechanism for video transmission in ieee 802.11 e wlans,”Wireless Personal Communications, vol. 85, no. 3, pp. 899–924, 2015.[80] M. Al Maqri, M. Othman, B. Ali, and Z. Hanapi, “Adaptive TXOPassignment for QoS support of video traffic in IEEE 802.11e networks,”in IEEE International RF and Microwave Conference (RFM), Dec 2013,pp. 144–149.[81] D. Skyrianoglou, N. Passas, and A. Salkintzis, “ARROW: AnEfficient Traffic Scheduling Algorithm for IEEE 802.11e HCCA,”IEEE Transactions on Wireless Communications, vol. 5, no. 12, pp.3558–3567, 2006.[82] Nikos Passas, Dimitris Skyrianoglou, and Panagiotis Mouziouras,“Prioritized support of different traffic classes in IEEE 802.11e wirelessLANs,” Computer Communications, vol. 29, no. 15, pp. 2867–2880,2006.[83] D.-Y. Lee, S.-R. Kim, and C.-W. Lee, “An Enhanced EDD QoS Schedulerfor IEEE 802.11e WLAN,” in Advances in Computational Science andEngineering, 2009, vol. 28, pp. 45–59.[84] Rami Haddad and M. McGarry, “Feed Forward Bandwidth Indication(FFBI): Cooperation for an accurate bandwidth forecast,” ComputerCommunications, vol. 35, no. 6, pp. 748–758, 2012.[85] D. Gao, J. Cai, and C. W. Chen, “Admission control based onrate-variance envelop for VBR traffic over IEEE 802.11e HCCAWLANs,” IEEE Transactions on Vehicular Technology, vol. 57, no. 3,pp. 1778–1788, 2008.[86] E. W. Knightly, “Second moment resource allocation in multi-servicenetworks,” in ACM SIGMETRICS Performance Evaluation Review,1997, pp. 181–191.[87] ——, “Enforceable quality of service guarantees for bursty trafficstreams,” in Proceeding IEEE INFOCOM ’98. Seventeenth Annual JointConference of the IEEE Computer and Communications Societies, vol. 2,1998, pp. 635–642.[88] W. F. Fan, D. Gao, D. H. Tsang, and B. Bensaou, “Admission controlfor variable bit rate traffic in IEEE 802.11e WLANs,” in The 13th IEEEWorkshop on Local and Metropolitan Area Networks, 2004, pp. 61–66.[89] C.-T. Chou, N. S. Shankar, and K. G. Shin, “Achieving per-stream QoSwith distributed airtime allocation and admission control in IEEE 802.11e wireless LANs,” in INFOCOM 2005. Proceedings IEEE 24th AnnualJoint Conference of the IEEE Computer and Communications Societies,vol. 3, 2005, pp. 1584–1595.[90] Q. Zhao and D. H. Tsang, “An equal-spacing-based design forQoS guarantee in IEEE 802.11e HCCA wireless networks,” IEEETransactions on Mobile Computing, vol. 7, no. 12, pp. 1474–1490, 2008.[91] S. Hantrakoon and A. Phonphoem, “Priority Based HCCA for IEEE802.11e,” in International Conference on Communications and MobileComputing, vol. 3, 2010, pp. 485–489.[92] M. O. Mohammed A. Al-Maqri, “An effcient hcca scheduler for videostreaming with qos support,” in The 8th International Conference onAmbient Systems, Networks and Technologies, 2017, pp. 1–8.[93] A. M. Mansoor, M. A. Al-Maqri, A. Q. Sabri, H. Jalab, A. W. A. Wahab,and W. kahtan Al-kopati, “A feedback-based admission control unit forqos provision of video transmission over wlans,” in IEEE 7th AnnualComputing and Communication Workshop and Conference (CCWC).IEEE, 2017, pp. 1–6.[94] “IEEE standard for information technology–local and metropolitan areanetworks–specific requirements–part 11: Wireless lan medium accesscontrol (mac) and physical layer (phy) specifications - amendment 8:Medium access control (mac) quality of service enhancements,” IEEEStd 802.11e-2005 (Amendment to IEEE Std 802.11, 1999 Edition (Reaff2003), pp. 1–212, Nov 2005.[95] Y.-J. Kim and Y.-J. Suh, “Adaptive polling MAC schemes for IEEE802.11 wireless LANs supporting voice-over-IP (VoIP) services,”Wireless Communications and Mobile Computing, vol. 4, no. 8, pp.903–916, 2004.[96] Xiyan Ma, Yanfeng Zhu, and Zhisheng Niu, “Dynamic pollingmanagement for QoS differentiation in IEEE 802.11e wireless LANs,” in10th Asia-Pacific Conference on Communications and 5th lnternatianalSymposium on Multi-Dimensional Mobile Communications, vol. 1,2004, pp. 152–156.[97] Byung-Seo Kim, Sung Won Kim, Yuguang Fang, and T. Wong,“Two-step multipolling MAC protocol for wireless LANs,” IEEE Journalon Selected Areas in Communications, vol. 23, no. 6, pp. 1276–1286,2005. [98] N. Ramos, D. Panigrahi, and S. Dey, “Dynamic adaptation policiesto improve quality of service of real-time multimedia applications inIEEE 802.11e WLAN Networks,” Wireless Networks, vol. 13, no. 4, pp.511–535, 2007.[99] J. Park, K. Cho, M. Choi, B. Lee, B. Lee, K. Kim, and K. Han, “A PollingScheme of TXOP Using Knapsack Algorithm in Wireless LAN,” inProceedings of the 8th WSEAS International Conference on EvolutionaryComputing, 2007, pp. 286–290.[100] Y.-C. Chen and H.-R. Yeh, “An Adaptive Polling Scheme SupportingAudio/Video Streaming in Wireless LANs,” in Proceedings of the 200812th IEEE International Workshop on Future Trends of DistributedComputing Systems, 2008, pp. 16–22.[101] Yong He and Xiaojun Ma, “Deterministic Backoff: Toward EfficientPolling for IEEE 802.11e HCCA in Wireless Home Networks,” IEEETransactions on Mobile Computing, vol. 10, no. 12, pp. 1726–1740,2011.[102] Chin-Wen Chou, K.-J. Lin, and Tsern-Huei Lee, “On efficientmultipolling with various service intervals for IEEE 802.11e WLANs,”in 7th International Wireless Communications and Mobile ComputingConference (IWCMC), 2011, pp. 1906–1911.[103] R. Viegas, L. Affonso, F. Vasques, P. Portugal, and R. Moraes,“Real-Time Industrial Communication over IEEE802.11e Wireless LocalArea Networks,” IEEE Latin America Transactions, vol. 10, no. 3, pp.1844–1849, 2012.[104] Li Feng, Jianqing Li, and Xiaodong Lin, “A New Delay Analysis forIEEE 802.11 PCF,” IEEE Transactions on Vehicular Technology, vol. 62,no. 8, pp. 4064–4069, 2013.[105] Zi-Tsan Chou, Cong-Qi Huang, and J. Chang, “QoS Provisioning forWireless LANs With Multi-Beam Access Point,” IEEE Transactions onMobile Computing, vol. 13, no. 9, pp. 2113–2127, 2014.[106] B. Zhang, M. Ma, C. Liu, and Y. Shu, “Performance improvementsof HCCA scheduling in V2R environments,” International Journal ofCommunication Systems, vol. 28, no. 5, pp. 861–872, 2015.[107] M. Yamane, S. Tagashira, and S. Fujita, “An Efficient Assignmentof Transmission Opportunity in QoS Guaranteed Wireless LAN,” inSeventh International Conference on Parallel and Distributed Computing,Applications and Technologies, 2006, pp. 105–108.[108] P. Ansel, Q. Ni, and T. Turletti, “FHCF: a simple and efficient schedulingscheme for IEEE 802.11e wireless LAN,” Journal Mobile Networks andApplications, vol. 11, no. 3, pp. 391–403, 2006.[109] Y. Choi, B. Lee, J. Pak, I. Lee, H. Lee, J. Yoon, and K. Han, “An AdaptiveTXOP Allocation in IEEE 802.11e WLANs,” in Proceedings of the 6thWSEAS International Conference on Electronics, Hardware, Wirelessand Optical Communications, 2007, pp. 187–192.[110] M. Rashid, E. Hossain, and V. Bhargava, “HCCA Scheduler Design forGuaranteed QoS in IEEE 802.11e Based WLANs,” in IEEE WirelessCommunications and Networking Conference (WCNC 2007), 2007, pp.1538–1543.[111] Kyung Mook Lim, Kang Yong Lee, Kyung Soo Kim, and Seong-SoonJoo, “Traffic Aware HCCA Scheduling for the IEEE 802.11e WirelessLAN,” in 2007 IEEE International Conference on Multimedia and Expo,2007, pp. 963–966.[112] Byung Joon Oh and Chang Wen Chen, “A cross-layer adaptation HCCAMAC for QoS-aware H.264 video communications over Wireless MeshNetworks,” in Proceedings of 2010 IEEE International Symposium onCircuits and Systems (ISCAS), 2010, pp. 2259–2262.[113] Kang Yong Lee, Kee Seong Cho, and Won Ryu, “Efficient QoSScheduling Algorithm for Multimedia Services in IEEE 802.11eWLAN,” in 2011 IEEE Vehicular Technology Conference, 2011, pp. 1–6.[114] A. Floros and N. Kanellopoulos, “A low complexity IEEE802.11escheduling scheme for efficient wireless delivery,” in 3rd InternationalSymposium on Communications, Control and Signal Processing, 2008,pp. 1–5.[115] Jing-Rong Hsieh and Tsern-Huei Lee, “Data rate estimation algorithmfor IEEE 802.11e HCCA scheduler,” International Journal of PervasiveComputing and Communications, vol. 3, no. 3, pp. 243–256, 2008.[116] X. Chu, “Provisioning of Parameterized Quality of Service in 802.11eBased Wireless Mesh Networks,” Mobile Networks and Applications,vol. 13, no. 1, pp. 6–18, 2008.[117] Deyun Gao, Jianfei Cai, and King Ngi Ngan, “Admission control in IEEE802.11e wireless LANs,” IEEE Network, vol. 19, no. 4, pp. 6–13, 2005.[118] Deyun Gao, Jianfei Cai, and L. Zhang, “Physical rate based admissioncontrol for HCCA in IEEE 802.11e WLANs,” in 19th International
VOLUME 0, 2018
17. A. Al-Maqri et al. : Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e
Conference on Advanced Information Networking and Applications,vol. 1, 2005, pp. 479–483.[119] M. van der Schaar, Y. Andreopoulos, and Zhiping Hu, “Optimizedscalable video streaming over IEEE 802.11 a/e HCCA wireless networksunder delay constraints,” IEEE Transactions on Mobile Computing,vol. 5, no. 6, pp. 755–768, 2006.[120] S. Jang and Y. Jang, “The Soft QoS-Aware Call Admission ControlScheme for HCCA in IEEE 802.11e,” in Information Networking.Advances in Data Communications and Wireless Networks, I. Chong andK. Kawahara, Eds. Springer Berlin Heidelberg, 2006, pp. 146–155.[121] G. Cecchetti, A. L. Ruscelli, and F. Checconi, “W-CBS: A SchedulingAlgorithm for Supporting QoS in IEEE 802.11e,” in The FourthInternational Conference on Heterogeneous Networking for Quality,Reliability, Security and Robustness & Workshops, 2007, pp. 1–7.[122] Qinglin Zhao and D. Tsang, “An Equal-Spacing-Based Design forQoS Guarantee in IEEE 802.11e HCCA Wireless Networks,” IEEETransactions on Mobile Computing, vol. 7, no. 12, pp. 1474–1490, 2008.[123] Zeng Ju ling, Xie bing, Zhou wen an, and Song Jun de, “Noticeof Retraction An improved admission control for HCCA in IEEE802.11e WLANS,” in 2008 11th IEEE International Conference onCommunication Technology, 2008, pp. 89–92.[124] F. Didi, H. Labiod, G. Pujolle, and M. Feham, “Physical rate andcontention window based admission control (PRCW) for 802.11WLANs,” in Computers and Communications (ISCC), 2010 IEEESymposium on, 2010, pp. 1–7.[125] N. K. Pali, M. Chawla, and J. Singhai, “RTS-AC: Admission ControlMethod for IEEE 802.11 e WLANs,” Mobile Ad-hoc Networks, pp. 8–12,2010.[126] Chie Dou and Chih-Wei Wu, “On the Effectiveness of Retransmissionin Preserving Frame Error Rate over IEEE 802.11e/a WLANs,” in2011 Third International Conference on Communications and MobileComputing, 2011, pp. 539–543.[127] J.-J. Huang, Y.-J. Liang, and C.-Y. Su, “Capacity enhancement for arate-variance-envelop-based admission control in IEEE 802.11e HCCAWLANs,” Wireless Networks, vol. 21, no. 7, pp. 2253–2261, 2015.[128] Y. Xiao, “IEEE 802.11e: QoS provisioning at the MAC layer,” WirelessCommunications, IEEE, vol. 11, no. 3, pp. 72–79, 2004.[129] Y. P. Fallah and H. Alnuweiri, “Hybrid polling and contention accessscheduling in IEEE 802.11e WLANs,” Journal of Parallel and DistributedComputing, vol. 67, no. 2, pp. 242–256, 2007.[130] Yaser Pourmohammadi Fallah and Hussein Alnuweiri, “Analysis oftemporal and throughput fair scheduling in multirate WLANs,” ComputerNetworks, vol. 52, no. 16, pp. 3169 – 3183, 2008.[131] Rongbo Zhu, Jiangqing Wang, and Maode Ma, “Intelligent MACmodel for traffic scheduling in IEEE 802.11e wireless LANs,” AppliedMathematics and Computation, vol. 205, no. 1, pp. 109–122, 2008.[132] M. Siddique, B.-L. Wenning, A. Timm Giel, C. Gorg, and M. Muhleisen,“Generic Spectrum Sharing Method Applied to IEEE 802.11e WLANs,”in Sixth Advanced International Conference on Telecommunications(AICT), 2010, pp. 57–63.[133] G. Boggia, P. Camarda, L. Grieco, and S. Mascolo, “Feedback-basedbandwidth allocation with call admission control for providing delayguarantees in IEEE 802.11e networks,” Computer Communications,vol. 28, no. 3, pp. 325 – 337, 2005.[134] Qiang Ni, “Performance analysis and enhancements for IEEE 802.11ewireless networks,” IEEE Network, vol. 19, no. 4, pp. 21–27, 2005.[135] H. Trsek, J. Jasperneite, and S. Karanam, “A Simulation Case Studyof the new IEEE 802.11e HCCA mechanism in Industrial WirelessNetworks,” in IEEE Conference on Emerging Technologies and FactoryAutomation, 2006, pp. 921–928.[136] S. Karanam, H. Trsek, and J. Jasperneite, “Potential of the HCCA schemedefined in IEEE802.11e for QoS enabled Industrial Wireless Networks,”in 2006 IEEE International Workshop on Factory CommunicationSystems, 2006, pp. 227–230.[137] M. Rashid, E. Hossain, and V. Bhargava, “Queueing Analysis of 802.11eHCCA with Variable Bit Rate Traffic,” in IEEE International Conferenceon Communications, vol. 10, 2006, pp. 4792–4798.[138] V. Siris and C. Courcoubetis, “Resource control for the EDCAmechanism in multi-rate IEEE 802.11e networks,” in InternationalSymposium on a World of Wireless, Mobile and Multimedia Networks,2006, pp. 419–428.[139] Z. Bin Muhamad Noh, T. Suzuki, and S. Tasaka, “Packet scheduling foruser-level QoS guarantee in audio-video transmission by IEEE 802.11e HCCA,” in TENCON 2007 - 2007 IEEE Region 10 Conference, 2007,pp. 1–4.[140] X. Perez Costa and D. Camps Mur, “IEEE 802.11E QoS and powersaving features overview and analysis of combined performance,” IEEEWireless Communications, vol. 17, no. 4, pp. 88–96, 2010.[141] Minseok Kim and Jong-Moon Chung, “Performance analysis of IEEE802.11e HCCA for V2I communications in WAVE networks,” in 53rdIEEE International Midwest Symposium on Circuits and Systems, 2010,pp. 328–331.[142] R. Ghazizadeh and P. Fan, “Queuing Analysis of HCCA for Multi-RateWireless LANs with Truncated ARQ Protocol,” Wireless PersonalCommunications, vol. 55, no. 4, pp. 607–630, 2010.[143] A. Lyakhov and M. Yakimov, “Analytical study of QoS-orientedmulticast in wireless networks,” EURASIP Journal on WirelessCommunications and Networking, vol. 2011, no. 1, pp. 1–13, 2011.[144] A. Pastrav, E. Puschita, and T. Palade, “HCCA support in IEEE 802.11networks QoS and QoE performance evaluation,” in 10th InternationalSymposium on Electronics and Telecommunications (ISETC), 2012, pp.139–142.[145] T. Lagkas, D. Stratogiannis, and P. Chatzimisios, “Modeling andperformance analysis of an alternative to IEEE 802.11e Hybrid ControlFunction,” Telecommunication Systems, vol. 52, no. 4, pp. 1961–1976,2013.[146] R. Ghazizadeh and Pingzhi Fan, “Queuing analysis for HCCA withadaptive modulation coding over wireless LANs,” in 11th IEEESingapore International Conference on Communication Systems, 2008,pp. 885–889.[147] W.-D. Lin and D.-J. Deng, “Service Differentiation in IEEE 802.11e HCF Access Method,” in Advances in Multimedia InformationProcessing, 2008, pp. 208–217.[148] A. Jansang, A. Phonphoem, and B. Paillassa, “Analytical Modelfor Expected Packet Delay Evaluation in IEEE 802.11e,” in WRIInternational Conference on Communications and Mobile Computing,vol. 2, 2009, pp. 344–348.
MOHAMMED A. AL-MAQRI received his BScdegree in Computer Science from AlmustanseriahUniversity - Iraq, in 2004. Then, he receivedhis MSc and Ph.D degrees in communicationtechnology and networks from Universiti PutraMalaysia in 2009 and 2016, respectively. Now,he is a Lecturer and Head of departmentof Information Technology in the Faculty ofComputer Science and Information Technology,Azal University for Human Development, Sana’a,Yemen. He has published a number of articles in high-impact factor scientificjournals. His research interests are in the field of high-speed TCP protocols,high-speed network, QoS, scheduling algorithms, admission control andwireless networks.
MOHAMED A. ALRSHAH (M’13–SM’17)received his BSc degree in Computer Sciencefrom Naser University - Libya, in June 2000.Then, he received his MSc and Ph.D degreesin communication technology and networks fromUniversiti Putra Malaysia (UPM) in May 2009and Feb 2017, respectively. Now, he is a SeniorLecturer in the Department of CommunicationTechnology and Networks, Faculty of ComputerScience and Information Technology, UniversitiPutra Malaysia (UPM). He has published a number of articles in high-impactfactor scientific journals. His research interests are in the field of high-speedTCP protocols, high-speed wired and wireless network, WSN, MANET,VANET, parallel and distributed algorithms, IoT and cloud computing. VOLUME 0, 2018 . A. Al-Maqri et al. : Review on QoS provisioning approaches for supporting video traffic in IEEE802.11e
MOHAMED OTHMAN (M’06–SM’18) receivedhis Ph.D from the Universiti Kebangsaan Malaysia(UKM) with distinction (Best Ph.D Thesis in 2000awarded by Sime Darby Malaysia and MalaysianMathematical Science Society). Now, he is aProfessor in the Department of CommunicationTechnology and Networks, Faculty of ComputerScience and Information Technology, UniversitiPutra Malaysia (UPM). He is also an associateresearcher at the Lab of Computational Scienceand Mathematical Physics, Institute of Mathematical Research (INSPEM),UPM. He published more than 160 International journals and 230proceeding papers. His main research interests are in the fields ofhigh-speed network, parallel and distributed algorithms, software definednetworking, network design and management, wireless network (MPDU-and MSDU-Frame aggregation, MAC layer, resource management, andtraffic monitoring) and scientific telegraph equation and modelling.