Documente Academic
Documente Profesional
Documente Cultură
A D VA N C E S I N W I R E L E S S V I D E O
CROSS-LAYER WIRELESS
MULTIMEDIA TRANSMISSION:
CHALLENGES, PRINCIPLES, AND NEW PARADIGMS
MIHAELA VAN DER SCHAAR, UNIVERSITY OF CALIFORNIA, DAVIS
SAI SHANKAR N, QUALCOMM INC., USA1
Input multimedia
(content characteristics, required
A
BSTRACT applications, such as videoconferencing, emer-
gency services, surveillance, telemedicine, remote
QoS, etc.) Wireless networks are poised to enable a teaching and training, augmented reality, and
variety of existing and emerging multimedia entertainment. However, existing wireless net-
streaming applications. As the use of wireless works provide only limited, time-varying quality
Optimize utility local area networks spreads beyond simple data of service (QoS) for delay-sensitive, bandwidth-
given constraints transfer to bandwidth-intense, delay-sensitive, intense, and loss-tolerant multimedia applications.
and loss-tolerant multimedia applications, Fortunately, multimedia applications can
addressing quality of service issues will become cope with a certain amount of packet losses
extremely important. Currently, a multitude of depending on the sequence characteristics and
Output
protection
(cross-layer adaptation and adaptation strategies exists in the
strategy) error concealment strategies available at the
different layers of the open systems interconnec- receiver. Consequently, unlike file transfers,
tion (OSI) stack. Hence, an in-depth under- real-time multimedia applications do not require
As the use of standing and comparative evaluation of these complete insulation from packet losses, but
strategies are necessary to effectively assess and rather that the application layer cooperate with
wireless local area enable the possible trade-offs in multimedia the lower layers to select the optimal wireless
networks spreads quality, power consumption, implementation
complexity, and spectrum utilization that are
transmission strategy that maximizes multimedia
performance.
beyond simple data provided by the various OSI layers. This further
NEED FOR CROSS-LAYER OPTIMIZATION
opens the question of cross-layer optimization
transfer to and its effectiveness in providing an improved In recent years the research focus has been to
solution with respect to the above trade-offs. In adapt existing algorithms and protocols for mul-
bandwidth-intense, this article we formalize the cross-layer problem, timedia compression and transmission to the
discuss its challenges, and present several possi- rapidly varying and often scarce resources of
delay-sensitive ble solutions. Moreover, we also discuss the wireless networks. However, these solutions
and loss tolerant impact the cross-layer optimization strategy
deployed at one station has on the multimedia
often do not provide adequate support for multi-
media applications in crowded wireless net-
multimedia performance of other stations. We introduce a works, when interference is high or stations are
new fairness concept for wireless multimedia sys- mobile. This is because the resource manage-
applications, tems that employs different cross-layer strate- ment, adaptation, and protection strategies avail-
gies, and show its advantages when compared to able in the lower layers of the stack the
addressing Quality of existing resource allocation mechanisms used in physical (PHY), medium access control (MAC),
Service (QoS) issues wireline communications. Finally, we propose a
new paradigm for wireless communications
and network/transport layers are optimized
without explicitly considering the specific charac-
will become based on coopetition, which allows wireless sta-
tions to harvest additional resources or free up
teristics of multimedia applications, and con-
versely, multimedia compression and streaming
extremely important. resources as well as optimally and dynamically algorithms do not consider the mechanisms pro-
adapt their cross-layer transmission strategies to vided by the lower layers for error protection,
improve multimedia quality and/or power con- scheduling, resource management, and so on.
1 The work was done sumption. This layered optimization leads to a simple
when the author was with independent implementation, but results in sub-
Philips Research USA. INTRODUCTION optimal multimedia (objective and/or perceptual
quality) performance. Alternatively, under
This work was supported Due to their flexible and low cost infrastructure, adverse conditions, wireless stations need to
in part by Intel IT wireless local area netwoks (LANs) are poised to optimally adapt their multimedia compression
Research. enable a variety of delay-sensitive multimedia and transmission strategies jointly across the
40 10
Node 8
Node 7
Node 8
9
35 Node 6
Node 7
8
30 Node 5
Node 6 7
Node 4
Effective goodput (Mb/s)
Node 2
4
15 Node 4
Node 1
3
Node 3
10
2
Node 2
5 Node 1
1
0 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Eb/N0 (dB) Eb/N0 (dB)
(a) (b)
n Figure 3. Optimal modulation scheme as a function of SNR and frame size decided by the MAC [12]: a) MSDU size 2000 octets; b)
MSDU size 200 octets.
40
pi i.e., pi = 1 . PHY 5
i =1, 2 PHY 3
35 PHY 1
Two simple strategies could be deployed for
transmission. In the first strategy, the wireless
PSNR (dB)
30
staion (WSTA) could send data only when the
channel is GOOD. In the second strategy, the 25
WSTA could use more power to transmit the
frame when the channel state is BAD to
20
decrease packet loss. To determine the optimal
transmission strategy, we consider the power
15
constraint as follows:
18 20 22 24 26 28 30
Pipiai Pmax, (7) SNR (dB)
22
Ti (t1, t2 ) i
, j = 1, 2,, N , (12)
18 T j (t1, t 2 ) j
14
where Ti and Tj represent the time allocated to
10 streams i and j, respectively.
AFS0 WFQ AFS WFQ AFS WFQ The advantages of the proposed air fair
Scenario 1 Scenario 2 Scenario 3 scheduler (AFS), [21], as opposed to the conven-
tional weighted fair queuing (WFQ), are high-
n Figure 5. The PSNR performance of multimedia for different fairness scenarios lighted in Fig. 5.
[20]. In the first scenario, the same cross-layer
strategies, resulting in the same transmission
rate, are deployed for both WSTAs. Thus, WFQ
Wi (t1, t 2 ) and AFS result in the same PSNR values. Con-
i r, sider the second scenario: WSTA1 experiences
t 2 t1 j (10)
more frame errors because of interference and
j fading. The packet loss rate (PLR) has
where r is the physical transmission rate or total increased, and it takes on average 50 percent
channel capacity. Thus, each multimedia flow i is more time to transmit a frame from WSTA1
guaranteed to have the throughput given by Eq. than from WSTA2. In conventional WFQ, this
10 regardless of the states of the queues and would mean that the start-of-service time of
frame arrivals of the other flows. However, the f rames in WSTA2 is deferred, resulting in QoS
advantages of using GPS, such as guaranteed violation and dropping of packets at the MAC
throughput and independent service, cannot be layer. This directly affects PSNR performance
preserved if the flows deploy different cross- as most of the higher-priority packets are
layer optimization, resulting in different trans- dropped for both WSTAs. Using AFS, the
mission rates. Depending on the channel stream between WSTA1 and the AP alone is
condition or their distance from the access point affected, yielding low PSNR, whereas WSTA2
(AP), WSTAs may choose different cross-layer is not affected because of WSTA1s channel
transmission strategies (PHY modes, retry limits, error condition. In scenario 3 WSTA1 moved
frame sizes, etc.) to ensure optimized multime- far away from the AP, and the cross-layer strat-
dia quality. Determining a fair share of resource egy switched the PHY mode to a more robust
among WSTAs in such a transmission scenario is modulation scheme. Since the physical trans-
a very challenging problem, because serving an mission rate of WSTA1 has dropped, it would
equal amount of traffic from individual stations take more time to transmit the frame, and the
deploying different strategies requires allocation same problem of deferred start of service hap-
of various amounts of air time and results in dif- pens for both WSTAs in WFQ. However, AFS
ferent impacts on the multimedia quality. isolates the channel and differential transmis-
sion rates to WSTA1, thus guaranteeing the
AIR OR TIME FAIRNESS multimedia performance.
We measure the total throughput degradation
due to WSTAs deploying different cross-layer MULTIMEDIA QUALITY FAIRNESS
strategies (e.g., different PHY rates) in the For multimedia applications, other fair schedul-
WLAN network. Given n WSTAs (with all sta- ing/allocation strategies could also be identified
tions having the same frame size), with besides equal time, such as equal multimedia
quality and guaranteed minimum quality. To ful-
8 fill a certain fairness criterion corresponding
ni ni = n
i =1 information (e.g., traffic characteristics, QoS
requirements) and resource exchange strategies
operating at, say, PHY mode i = (1, , 8), the for wireless multimedia are necessary.
throughput degradation can be determined as To understand the potential impact on mul-
[20] timedia quality due to different resource
1 exchanges and corresponding cross-layer adap-
Throughput = . tation, let us consider the example of multime-
1 8 n
j
(11) dia transmission over an IEEE 802.11e network.
n i =1 R j To fully utilize the features provided by the
MAC protocol for multimedia transmission, we
WSTAs having different transmission rates Rj propose to use the available application layer
due to the different PHY modes or other information to partition multimedia streams
deployed cross-layer optimization strategies into subflows with different priorities, delay
cause this unfairness. To solve this problem, we bounds, retry limits, and packet sizes. A base
propose the concept of time fairness. In this con- quality subflow can be admitted using an admis-
NEW COOPETITION PARADIGM FOR erning wireless systems more effectively, thereby
resulting in improved performance for multime-
WIRELESS MULTIMEDIA dia applications.
As discussed previously, wireless devices current- The costs associated with the resource
ly operate in a non-collaborative manner that exchange can be quantified in terms of the
limits their performance and overall wireless sys- degradation in multimedia quality, increased
tem performance, as competing stations do not delay, or power consumption [23]. For instance,
always effectively exploit available resources. if the resulting multimedia quality Q after the
Consequently, to improve the performance of resource exchange is positively impacted, or Q is
wireless multimedia applications, we discuss a above a certain maximum quality T1 (e.g., above
new paradigm that fundamentally changes the 40 dB), the WSTA will contribute its resources.
non-collaborative way in which WSTAs currently Alternatively, if Q is negatively impacted and
interact by allowing them to exchange informa- below T 1, a penalty will be associated with any
tion and distribute resources. The proposed resource exchange. If Q is below a minimum
paradigm was inspired by a relatively new and quality T 2 (e.g., below 28 dB), the WSTA will
successful economics concept known as coopeti- stop contributing resources.
tion [22], which suggests that a judicious mixture Our preliminary results in [24] have shown
of competition and cooperation is often advanta- that coopetition results in an improved number
geous in competitive environments. When of satisfied users (i.e., station satisfying their
applied to wireless multimedia systems, coopeti- minimal quality requirements) compared to de
tion fundamentally changes the passive way sta- facto allocation of resources. We have designed
tions currently adapt their transmission strategies different coopetition strategies [24] that con-
to match available wireless and power resources, verge to distinct Nash equilibriums depending
by enabling them to proactively influence the on the channel conditions, multimedia applica-
wireless systems dynamics through resource and tion characteristics, resource exchange policies,
information exchange. and so on, resulting in different cost-benefit
For example, two WSTAs experiencing a high trade-offs for the participating WSTAs. The
PLR over a channel with a high contention level design of optimal coopetition strategies together
can collaboratively decide to reduce their retry with cross-layer optimized design constitutes a
limit or adapt their contention parameters to vast topic of further research for improving
reduce contention and thus improve their overall future wireless multimedia systems.
multimedia performance and power consump-
tion. CONCLUSIONS
To allow coopetition, we propose a new way
of architecting the wireless multimedia commu- The previously described cross-layer optimized
nication system by jointly optimizing the proto- wireless multimedia paradigm is only recently
col stack at each station and the resource emerging, and a variety of research topics still
exchanges among stations. In the proposed need to be addressed. Realistic integrated mod-
paradigm, information about resources and con- els for the delay, multimedia quality, and con-
straints (e.g., QoS requirements, multimedia sumed power of various transmission strategies/
traffic characteristics, experienced channel con- protocols need to be developed. Moreover, the
ditions) of the various stations can be dissemi- benefits in terms of multimedia quality of
nated to all network members (stations), and employing a cross-layer optimized framework for
used as available optimization criteria for their different multimedia applications with different
own communication subsystem. The proposed delay sensitivities and loss tolerances still need
coopetition paradigm is a superset of proposed to be quantified. We have also identified a new
fairness concepts that can be deployed for gov- paradigm for wireless multimedia transmission