Impact of the IEEE 802.11aa Intra-AC Prioritization on QoE for H.264

Compression

Katarzyna Kosek-Szott, Marek Natkaniec, Lukasz Prasnal and Lucjan Janowski

AGH University of Science and Technology, al. Mickiewicza 30, 30-059 Krakow, Poland

Keywords:

IEEE 802.11aa, CBR, Experiment, H.264, intra-AC Prioritization, Mapping, VBR, Video Streams.

Abstract:

Recently, a new 802.11aa amendment has been released. Its main goal is to increase quality of service support

for audio-video trafﬁc streams. Among others, it deﬁnes the intra-access category prioritization mechanism

coupled with appropriate trafﬁc selection procedures, in order to increase the granularity of trafﬁc prioritization

in comparison to the currently used EDCA. This paper presents the preliminary study of the possible impact

of the new feature on quality of experience (QoE) in case of H.264 video streams. The obtained results show

that IEEE 802.11aa and 802.1Q parameters should be tuned to efﬁciently use the available bandwidth.

1 INTRODUCTION

The transmission of audio-video trafﬁc streams over

IEEE 802.11 networks has recently become very pop-

ular. As a result, standardization bodies try to im-

prove the effectiveness and Quality of Service (QoS)

support in the IEEE 802.11 standard (IEEE 802.11,

2012) by adding new QoS features and increasing

theoretical network throughput in the form of new

amendments. The ﬁrst QoS successor to the orig-

inal distributed coordination function (DCF) is the

enhanced distributed channel access (EDCA), which

maps trafﬁc streams to four independent access cat-

egories (ACs): voice (VO), video (VI), best effort

(BE), and background (BK). The granularity of traf-

ﬁc prioritization deﬁned by EDCA is rather limited

because differentiation of trafﬁc streams within a sin-

gle AC is not supported. The second QoS successor

of DCF is the recently proposed 802.11aa amendment

(IEEE 802.11aa, 2012), which deﬁnes a number of

mechanisms to improve audio-video streaming over

wireless local area networks (WLANs) in comparison

to EDCA (Kosek-Szott et al., 2012) and introduces

the intra-ac prioritization.

Since the appearance of EDCA a number of pa-

pers addressed the problem of the possibility of map-

ping of different priority video frames to different

ACs for the H.264 compression (Choudhry and Kim,

2005), (Koo et al., 2006), (Ksentini et al., 2006), (Pli-

akas et al., 2007), (Birlik et al., 2007), (Milani et al.,

2008), (Chen et al., 2008), (MacKenzie et al., 2009),

(Birlik et al., 2009), (Politis et al., 2011), (Debnath,

2012), (Yoon et al., 2012), (Wang and Liu, 2012),

(Yao et al., 2013). This proves the importance of

the described research area. However, only several

of them concentrate both on video transmission and

Quality of Experience (QoE) (Debnath, 2012) (Wang

and Liu, 2012), which is an important subjective mea-

sure of a customer’s satisfaction with a service (e.g.,

video streaming). There is also one paper in which the

authors propose an intra-access category trafﬁc priori-

tization (Sutinen and Huusko, 2011) coupled with the

adaption of H.264 streams in the MAC layer. How-

ever, in the literature there are no papers describing

802.11aa in the context of QoE.

In this position paper, we focus on the 802.11aa

intra-AC prioritization of H.264 video frames using

two transmission queues: primary VI and alternate

VI coupled with a credit-based transmission selection

algorithm (cf. Section 2). We illustrate how the in-

creased granularity of trafﬁc prioritization may im-

pact the QoE in WLANs (cf. Section 3 and 3.1).

2 INTRA-AC PRIORITIZATION

AND CREDIT-BASED SHAPING

The intra-AC prioritization mechanism extends the

operation of legacy EDCA by deﬁning alternate MAC

transmission queues for the VO and VI ACs to ob-

tain a ﬁner-grained prioritization between individual

audio and video trafﬁc streams. As a result, 802.11aa

deﬁnes six transmission queues: twoVO (primary VO

Kosek-Szott K., Natkaniec M., Prasnal L. and Janowski L..

Impact of the IEEE 802.11aa Intra-AC Prioritization on QoE for H.264 Compression.

DOI: 10.5220/0005118800650070

In Proceedings of the 11th International Conference on Wireless Information Networks and Systems (WINSYS-2014), pages 65-70

ISBN: 978-989-758-047-5

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

Classification to transmit queue and access category

VO VI BE BK

Alternate

(A_VO)

Primary

(VO)

Primary

(VI)

Alternate

(A_VI)

BE BK

Frame transmission attempt

Transmit

queues

EDCA functions

with internal

collision resolution

Highest priority Lowest priority

IEEE 802.1Q

scheduling

MSDU (UP)

Figure 1: Trafﬁc prioritization in 802.11aa.

and alternate A

VO), two VI (primary VI and alter-

nate A VI), BE, and BK (Fig. 1). The six queues

are derived from the 802.1D user priorities (UPs)

(IEEE 802.1d, 2004). Frames belonging to compet-

ing queues within an AC are selected using an appro-

priate transmission selection algorithm deﬁned by the

802.1Q standard (IEEE 802.1q, 2011): strict prior-

ity algorithm (SPA) or credit-based shaper algorithm

(CBSA)

. Importantly, the transmission selection al-

gorithm must be conﬁgured so that frames belonging

to the queue with the higher UP are selected with a

higher probability than from the lower priority queue.

Having been scheduled, frames are mapped to four

independent EDCA functions and the actual frame

transmission is organized using the standard 802.11

transmission procedures.

In CBSA, frame selection is based on an internal

credit parameter. A frame belonging to a given queue

is selected only if (i) for the primary transmit queue

credit is non-positive or (ii) for the alternate transmit

queue credit is either positive or when credit is equal

to zero and the primary queue is empty. The value of

credit is calculated based on two external parameters:

portTransmitRate—the transmission rate, in bits per

second, supported by the underlying MAC service,

and idleSlope—the rate of change of credit, in bits per

second, when the value of credit increases. The lat-

ter determines the maximum portion of portTransmit-

Rate available for the transmission of frames stored in

the alternate queue. Additionally, sendSlope, an inter-

nal parameter, determines the rate of credit change, in

bits per second, when the value of credit decreases:

sendSlope = idleSlope− portTransmitRate. (1)

The operation of CBSA is illustrated in Fig. 2.

The increase of credit occurs with a rate of idleSlope

SPA is the default transmission selection algorithm

which gives absolute priority to the higher priority queue.

CBSA is an optional algorithm which allows ﬂexible band-

width allocation. We selected CBSA in our tests, as more

powerful.

(i) during the transmission of a frame from the pri-

mary queue and (ii) when there is no transmission

while credit is negative. Conversely, credit is de-

creased with a rate of sendSlope during the transmis-

sion of a frame from the alternate queue. Addition-

ally, if credit is positive and the alternate queue is

empty then credit is reset to zero.

Apart from the main parameters of the CBSA al-

gorithm the following auxiliary values are deﬁned in

802.1Q: loCredit—the minimum value that can be ac-

cumulated in the credit parameter:

loCredit = maxFrameSize×

sendSlope

portTransmitRate

(2)

and hiCredit—the maximum value that can be accu-

mulated in the credit parameter:

hiCredit = maxInterferenceSize×

idleSlope

portTransmitRate

(3)

where maxFrameSize is the maximum size of a frame

that can be transmitted and maxInterferenceSize is the

maximum size of a burst of trafﬁc that can delay a

frame transmission.

hiCredit

loCredit

Increasing

Credit

Increasing

Time

Number of

Queued Frames

Frame C

Frame B

Frame A

Transmitted

Data

Conflicting

Frames

Fr. A

Frame B

Frame C

ACK

Fr. A

ACK

Fr. B

Fr. C

ACK

Frame C

Frame B

Frame C

delayed until

credit reaches

zero

Figure 2: CBSA operation for three frames (A, B, and C)

accumulated in the alternate queue during the transmission

of conﬂicting trafﬁc. Backoff time was named BF in the

ﬁgure.

3 EXPERIMENT SETTINGS

In order to analyze the impact of mapping of H.264

frames to different 802.11aa VI queues on QoE we

performed subjective tests. The tests were per-

formed over one week in June 2013 at the Depart-

ment of Telecommunications at AGH University of

Science and Technology in Poland. We selected eight

source sequences with different number of details and

WINSYS2014-InternationalConferenceonWirelessInformationNetworksandSystems

with different movement dynamics (media.xiph.org,

2014): (1) for the birds, (2) crowd run, (3) moss for-

est, (4) ducks take off, (5) speed bag, (6) park joy,

(7) controlled burn, and (8) old town cross. Each se-

quence was 10 seconds long. All video sequences

were encoded using ffmpeg converter integrated with

libx264 encoder. In predominant number of cases,

the native resolution of the original video traces (me-

dia.xiph.org, 2014) was 1920x1080 (full HD), how-

ever all video sequences were resized during the en-

coding process to 1280x720 (HD), which is better

supported by slower mobile devices. The frame rate

was equal to 50 fps. Each sequence was encoded three

times: to obtain ﬁxed quality scale (VBR), 9 Mbps

CBR, and 12 Mbps CBR. Two-pass encoding option

was set to efﬁciently use the assumed bitrates. The

maximum size of GOP was set to 18 and it determined

the maximum distance between I-frames. Every P

frame was separated by two B frames with hierarchi-

cal coding: B-frame (reference frame) and b-frame

(non-referenceframe). Therefore, the maximum GOP

structure was IbBPbBPbBPbBPbBPbP.

The analyzed test sequences were evaluated by 24

testers who varied in age (there were eight testers

in each group: 18 − 26 years, 27 − 38 years, and

39− 50 years) and sex (there were 14 females and 10

males). The chosen test sequences were displayed in

the center of a 15.6 LCD monitor with a resolution of

1600x900 pixels. Prior to the QoE tests, the testers’

sight was evaluated using a visual acuity and a color

blindness tests. The sequences were played out in the

VLC-2.0.7 multimedia player

(VLC, 2014) in a ran-

dom order. The testers scored each sequence using the

MOS scale. Each experiment started with three train-

ing sequences in order to familiarize the tester with

the speciﬁcity of the test.

3.1 Test Sequences

The ns-3.17 simulator (www.nsnam.org, 2014) was

selected to prepare the tested video sequences with

appropriate mapping of different video frames. We

have implemented the complete 802.11aa intra-AC

prioritization feature together with the CBSA trafﬁc

selection procedure in the the simulator. Moreover

the EvalVid module adopted for ns-3.17 was added

(on the basis of Evalvid module for ns-3.13 developed

by GERCOM (gercom.ufpa.br,2014)).

A Video Quality Evaluation Tool-set 2.7 was used

for the purposes of damaging video ﬁles in accor-

dance to simulation results and obtaining transmis-

sion statistics like frame losses (Klaue et al., 2003),

We selected VLC as one of the most commonly used

free multimedia players.

Original test

sequence

Encoder

H.264 bit stream

ns-3 simulator

Input trace

(MP4Box, mp4trace)

Output trace

Video reconstruction

(etmp4)

ns-3 simulation script with different

mappings of video frames to 802.1D

UPs:

Scenario A: {I, P + B, b} = {4, 4, 4}

Scenario B: {I, P + B, b} = {5, 4, 4}

Scenario C: {I, P + B, b} = {5, 5, 4}

Scenario D: {I, P + B, b} = {5, 5, 5}

Figure 3: Preparation of video sequences for QoE tests.

(EvalVid, 2014). In order to simulate transport with

RTP, at the ﬁrst step the hint track for the input video

ﬁles was added with the use of the MP4Box - an appli-

cation being a part of GPAC multimedia framework

(gpac.wp.mines-telecom.fr, 2014). This preparation

was done for the transmission with MTU size set to

1500 bytes. Next, the mp4trace application from the

EvalVid was used to obtain input trace ﬁles for the

EvalVid module implemented in ns-3. However, be-

cause mp4trace in EvalVid-2.7 was unable to recog-

nize B frames treating them like P frames, a slight

modiﬁcation was added. The result was that mp4trace

was able to recognize three kinds of frames: 1) intra-

coded frames (I frames), 2) not-intra-coded frames

being reference frames (P and B frames) and 3) non-

reference b frames. In consequence our frame cat-

egorization was performed basing not only on their

prediction category but also on the frames’ joint de-

pendence.

During the simulations the ns-3 EvalVid module

produced output trace ﬁles which were then passed

to etmp4 application calculating evaluation results. In

this manner we obatained not only statistical informa-

tion about frame losses but also output video ﬁles with

damages which were then used for QoE examination.

Importantly, etmp4 did not perform slice reconstruc-

tion and, in consequence, a whole video frame was

treated as lost in case of any of its data loss.

The whole process of video sequences preparation

is shownin Fig.3. Its main steps are the following: en-

coding, trace ﬁle preparation, simulation of different

scenarios and, ﬁnally, video reconstruction according

to the simulation output trace ﬁle.

ImpactoftheIEEE802.11aaIntra-ACPrioritizationonQoEforH.264Compression

Table 1: FLRs of sequences 1-8 for scenarios A-D and for variable rates (CBR 9 Mpbs, CBR 12 Mbps, and VBR).

Scen. Rate Seq. No.

FLR [%]

Scen. Rate Seq. No.

FLR [%]

I P + B b All I P + B b All

1 0 7.87 1.82 5.41

1 0 0 0 0

Mbps

2 0 8.82 10.24 8.8

Mbps

2 0 0 0 0

3 0 12.34 1.2 7.95 3 42.86 0.32 0 2.58

4 0 2.61 5.42 3.4 4 0 0 0 0

5 0 8.5 10.24 8.6 5 20.69 0 0 1.2

6 0 6.86 9.04 7.2 6 0 0 0 0

7 0 9.48 16.48 11.21 7 39.29 0 0 2.2

8 0 5.99 11.93 7.47 8 0 0 0 0

1 0 16.82 2.84 11.21

1 0 0 0 0

Mbps

2 0 11.11 16.87 12.4

Mbps

2 21.43 0 0 1.2

3 0 14.61 5.99 10.93 3 13.16 0 0 0.91

4 0 16.99 39.76 23.6 4 42.86 10.06 1.2 8.95

5 0 18.86 19.89 17.85 5 0 1.2 0 0.73

6 0 8.82 19.28 11.8 6 0 0 0 0

7 5.26 13.73 5.11 10.38 7 12.5 2.75 0 2.43

8 0 9.15 16.87 11.2 8 0 0 0 0

VBR

1 0 8.5 0 5.2

VBR

1 7.14 0 0 0.4

2 0 0 0 0 2 0 0 0 0

3 0 0 0 0 3 14.29 6.86 0 5

4 0 38.56 58.43 43 4 0 0 0 0

5 0 8.44 7.78 7.75 5 34.48 0 0 2

6 0 17.32 63.25 31.6 6 0 0.65 0 0.4

7 0 7.87 0.61 5.01 7 39.29 0 0 2.2

8 0 39.87 74.1 49 8 0 0 0 0

1 0 0 0 0

1 0 0.32 0 0.2

Mbps

2 42.86 1.3 0 3.18

Mbps

2 0 0 0 0

3 0 3.67 9.66 5.42 3 0 0 0 0

4 0 0 0 0 4 0 0 0 0

5 24.14 0.66 0.61 2 5 0 0 0 0

6 28.57 0 0 1.6 6 0 0 0 0

7 0 0.6 2.27 1.09 7 0 0 0 0

8 0 0 0 0 8 0 0 0 0

1 6.25 10.7 2.84 7.85

1 0 10.39 3.59 7.55

Mbps

2 0 7.78 13.07 8.93

Mbps

2 0 0 0 0

3 0 8.82 24.1 13.4 3 0 0 0 0

4 0 0 1.2 0.4 4 0 0 0 0

5 0 0.33 2.41 1 5 0 0 0 0

6 0 0 1.81 0.6 6 0 0 0 0

7 13.16 2.99 3.41 3.83 7 0 0 0 0

8 50 12.66 10.18 13.92 8 0 0 0 0

VBR

1 0 8.5 49.4 21.6

VBR

1 0 0 0 0

2 17.86 0 0 1 2 0 0 0 0

3 0 0 0 0 3 0 0 0 0

4 37.93 0 0 2.2 4 0 0 0 0

5 0 33.33 57.83 39.6 5 0 0 0 0

6 0 0 0 0 6 0 0.33 0.6 0.4

7 0 0 0 0 7 0 0 0 0

8 14.29 30.72 63.86 40.8 8 0 0 0 0

3.2 Parameters and Scenarios

The simulation parameters are presented in Table 2.

802.11a was used at the PHY layer. The wireless

channel introduced no errors. The standard EDCA

parameters were used at the MAC layer. The idle-

Slope of CBSA was set to 25%. Analyzed scenarios

are presented in Fig. 3. Frame loss ratios (FLRs) of

different video frame types in all test sequences are

presented in Table 1.

4 SUBJECTIVE TEST RESULTS

We performed a post-experiment inspection of the

subjective results. The rejection criteria (correlation

Table 2: Simulation settings.

Parameter Value Parameter Value

PHY layer 802.11a CW

min

Data rate 54 Mbps CW

max

Basic rate 6 Mbps AIFSN 2

RTS/CTS Turned off Queue size 400 frames

Mode Ad-hoc MSDU Lifetime 100 ms

SIFS 16 µs Slot time 9 µs

Preamble 16 µs PLCP header 4 µs

MTU 1500 B No. of stations 1

IdleSlope 25% No. of streams 1

coefﬁcient R

< 0.75 ) veriﬁed the level of correlation

of the scores of one tester according to the mean score

of all the testers over the entire experiment. It ap-

WINSYS2014-InternationalConferenceonWirelessInformationNetworksandSystems

700 800 900 1000 1100 1200 1300 1400 1500

Kruskal–Wallis test

Ranks

Sequence no

Figure 4: Veriﬁcation of test sequences.

900 1000 1100 1200 1300 1400 1500 1600 1700

555

554

544

444

Mean rank

Mapping type

(a)

600 800 1000 1200 1400 1600 1800

3555

3554

3544

3444

2555

2554

2544

2444

1555

1554

1544

1444

Kruskal–Wallis test

Rank

Sequence type/mapping type

(b)

Figure 5: Subjective test results: mean ranks for different

mapping types (a), mean ranks for different mapping types

for different bit rates (1 – CBR, 12 Mbps, 2 – CBR, 9 Mbps,

3 – VBR) and mapping types (the last three positions in the

Y axis) (b).

peared that all testers appropriately ranked the viewed

sentences. Fig. 4 additionally shows mean ranks de-

pending on the source sequence number. It proves

that the test sentences were appropriately varied.

Fig. 5(a) shows mean ranks for different mapping

types. As can be seen only the mapping proposed in

Scenario D ({I, P + B, b}={5, 5, 5}) is considerably

better from the three others. This means that static as-

signment of different priority H.264 video frames into

both primary and alternate transmission queues is not

satisfactory and worsens the overall QoE. Addition-

ally, Fig. 5(b) shows mean ranks of different mapping

types for different bit rates (1 – CBR, 12 Mbps, 2 –

CBR, 9 Mbps, 3 – VBR). The results are in line with

Fig. 5(a).

5 CONCLUSION

In this work-in-progress paper we show that the static

setting of the CBSA parameters is not appropriate to

improve the QoE in WLANs. This is because the

main objective of the CBSA algorithm is to reserve

some part of throughput for the alternative trafﬁc and

to simultaneously limit the same trafﬁc to the spec-

iﬁed level. With our default settings, the 25% of

throughput was reserved for the alternative queue and

thus the primary one has guaranteed 75% of through-

put. At the same time, the analyzed static mapping

was producing two streams of various bitrates depen-

dent on the selected scenario and the examined video

sequence.

In case of the CBR sequences, the prioritized I-

frames were huge and could generate trafﬁc overcom-

ing guaranteed throughput. At the same time alterna-

tive trafﬁc was smaller than the reserved 25% and thus

all of it was protected. In fact, the analyzed mecha-

nism gives reverse results to the original intentions -

small frames were transmitted successfully while the

most important I-frames were damaged because they

were not able to use the whole throughput when the

alternative queue was not empty. On the other hand,

in some other cases the amount of the limited through-

put could not be satisfactory and some frames were

dropped even if the trafﬁc in the prioritized queue did

not consume all of its guaranteed throughput.

The conclusion is that the CBSA settings should

be modiﬁed dynamically to adapt to current demands,

for example the idleSlope parameter should be in-

creased when the trafﬁc in the primary queue is small

enough. Another solution is to design a more intel-

ligent prioritization scheme. For example it could

resign from assigning the whole frame to the queue

with the limited bandwidth if there is no possibility to

transmit all of its data in time lower than the MSDU

Lifetime Limit. Therefore, as future work we envi-

sion designing self-conﬁguration mechanisms which

will allow appropriate setting of the parameters char-

acteristic to the CBSA algorithm in order to improve

the QoE in future WLANs and relieve network ad-

ministrators from complicated management.

ImpactoftheIEEE802.11aaIntra-ACPrioritizationonQoEforH.264Compression

ACKNOWLEDGEMENTS

This work has been supported by the AGH Univer-

sity of Science and Technology under contracts no.

15.11.230.052 and 11.11.230.018.

REFERENCES

Birlik, F., Ercetin, O., and Gurbuz, O. (2007). Prioritized

video streaming in wireless mesh networks. In Proc.

WoWMoM 2007, pages 1–3. IEEE.

Birlik, F., Gurbuz, O., and Ercetin, O. (2009). IPTVhome

networking via 802.11 wireless mesh networks: an

implementation experience. Consumer Electronics,

IEEE Transactions on, 55(3):1192–1199.

Chen, H.-L., Lee, P.-C., and Hu, S.-H. (2008). Im-

proving scalable video transmission over IEEE

802.11e through a cross-layer architecture. In Proc.

ICWMC’08, pages 241–246. IEEE.

Choudhry, U. I. and Kim, J. (2005). Performance evalua-

tion of H.264 mapping strategies over IEEE 802.11e

WLAN for robust video streaming. In Proc. PCM

2005, pages 818–829. Springer.

Debnath, T. (2012). A Novel QoS-Aware MPEG-4 Video

Delivery Algorithm over the Lossy IEEE 802.11

WLANs to Improve the Video Quality. PhD thesis,

School of Electronic and Communications Engineer-

ing, Dublin Institute of Technology.

EvalVid (2014). A Video Quality Evaluation Tool-set.

http://www2.tkn.tu-berlin.de/research/evalvid/.

gercom.ufpa.br (2014). NS3.13 + Evalvid module devel-

oped by GERCOM. http://gercom.ufpa.br/.

gpac.wp.mines-telecom.fr (2014). Multimedia Open

Source Project. http://gpac.wp.mines-telecom.fr/.

IEEE 802.11 (2012). Wireless LAN Medium Access Con-

trol (MAC) and Physical Layer (PHY) Speciﬁcations,

March 2012.

IEEE 802.11aa (2012). Wireless LAN Medium Access

Control (MAC) and Physical Layer (PHY) Speciﬁca-

tions Amendment 2: MAC Enhancements for Robust

Audio Video Streaming.

IEEE 802.1d (2004). IEEE 802.1D-2004 – IEEE Standard

for Local and metropolitan area networks: Media Ac-

cess Control (MAC) Bridges.

IEEE 802.1q (2011). IEEE 802.1Q-2011 — IEEE Stan-

dard for Local and metropolitan area networks–Media

Access Control (MAC) Bridges and Virtual Bridged

Local Area Networks.

Klaue, J., Rathke, B., and Wolisz, A. (2003). Evalvid–a

framework for video transmission and quality evalu-

ation. In Computer Performance Evaluation. Mod-

elling Techniques and Tools, pages 255–272. Springer.

Koo, S.-Y., Kim, B.-K., and Kim, Y.-T. (2006). Man-

agement of IEEE 802.11e wireless LAN for realtime

qos-guaranteed teleconference service with differen-

tiated H.264 video transmission. In Autonomic Man-

agement of Mobile Multimedia Services, pages 25–36.

Springer.

Kosek-Szott, K., Krasilov, A., Lyakhov, A., Natkaniec,

M., Safonov, A., Szott, S., and Tinnirello, I. (2012).

What‘s New for QoS in IEEE 802.11. IEEE Network,

27(6):95–104.

Ksentini, A., Naimi, M., and Gu´eroui, A.(2006). Toward an

improvement of H.264 video transmission over IEEE

802.11e through a cross-layer architecture. Communi-

cations Magazine, IEEE, 44(1):107–114.

MacKenzie, R., Hands, D., and O’Farrell, T. (2009). Ef-

fectiveness of H.264 error resilience techniques in

802.11e WLANs. In Proc. WCNC 2009, pages 1–6.

IEEE.

media.xiph.org (2014). Xiph.Org Foundation, Video Test

Sequences. http://media.xiph.org/video/derf/y4m/.

Milani, S., Calvagno, G., Bernardini, R., and Zontone, P.

(2008). Cross-layer joint optimization of FEC channel

codes and multiple description coding for video deliv-

ery over IEEE 802.11e Links. In Proc. NGMAST’08,

pages 472–478. IEEE.

Pliakas, T., Kormentzas, G., and Tsekeridou, S. (2007).

Joint scalable video coding and packet prioritization

for video streaming over IP/802.11e heterogeneous

networks. In Proc. international conference on Mo-

bile multimedia communications, page 31. ICST.

Politis, A., Mavridis, I., Manitsaris, A., and Hilas, C.

(2011). X-EDCA: A cross-layer MAC-centric mech-

anism for efﬁcient multimedia transmission in con-

gested IEEE 802.11e infrastructure networks. In Proc.

IWCMC 2011, pages 1724–1730. IEEE.

Sutinen, T. and Huusko, J. (2011). MAC layer QoS archi-

tecture for optimized and fair transmission of scalable

video. In CCNC 2011, pages 421–425. IEEE.

VLC (2014). Media player. http://www.videolan.org/vlc/.

Wang, H. and Liu, G. (2012). Priority and delay aware

packet management framework for real-time video

transport over 802.11e WLANs. Multimedia Tools

and Applications, pages 1–21.

www.nsnam.org (2014). Network simulator ns-3.

http://www.nsnam.org/.

Yao, X.-w., Wang, W.-l., Yang, S.-h., and Chen, S.-

y. (2013). PABM-EDCF: parameter adaptive bi-

directional mapping mechanism for video transmis-

sion over WSNs. Multimedia tools and applications,

63(3):809–831.

Yoon, Y., Kim, M., Lee, S., Lee, B., Hyun, S. J., and

Lee, K. (2012). Performance analysis of H.264/AVC,

H.264/SVC, and VP8 over IEEE 802.11 wireless net-

works. In IEEE Symposium on Computers and Com-

munications, 2012, pages 151–156. IEEE.

WINSYS2014-InternationalConferenceonWirelessInformationNetworksandSystems