5.6 Adaptation by Cross Layer Design


5.6 Adaptation by Cross Layer Design

An important aspect of wireless networks is dynamic behavior. The conventional protocol structure is inflexible as various protocol layers can only communicate in a strict manner. In such a case, the layers are designed to operate under the worst conditions, rather than adapting to changing conditions. This leads to inefficient use of spectrum and energy.

Adaptation represents the ability of network protocols and applications to observe and respond to the channel variation. Central to adaptation is the concept of cross layer design. [33], [34] Cross layer design for the three key layers in the overall protocol stack (i.e., application layer, transport layer, and network and link layer) are reviewed in this section. An example framework is illustrated in Figure 5.3 in terms of streaming video over wireline-to-wireless networks.

click to expand
Figure 5.3: A video streaming architecture using cross layer design.

5.6.1 Application Transmission Adaptation

Application Transmission Adaptation refers to the application's capability to adjust its behavior to changing network and channel characteristics. Wireless networks often have to deal with adverse conditions where handoffs, deep fading, and bad carrier signals result in a high rate of packet losses. Only adaptive applications can cope with these challenging circumstances. For multimedia delivery, a media server can track packet losses and adjust media source rate accordingly. [35], [36], [37], [38], [39] To reduce information loss, the media server can employ packet FEC coding and UEP, as described in Section 5.4.

Whereas this level of adaptation is system independent and application specific, an application is able to reconfigure itself accurately only if it identifies the underlying network and channel variations.

5.6.2 Transport Layer Transmission Adaptation

Instead of application layer adaptation, the adaptation can be shifted to the underlying transport layer, making it transparent to the application layer, so that applications originally developed for wireline networks remain intact. One drawback of this level of adaptation is that it is impossible to implement a complete adaptation if part of it is application specific.

The protocol should differentiate various packet loss patterns (i.e., packet losses due to network congestion or from channel errors), [40], [41], [42] and invoke congestion control and rate adaptation accordingly. Several cross layer approaches, such as EBSN, [43] snoop, [44] and freeze TCP, [45] have been proposed as TCP alternatives to distinguish congestion loss from noncongestion loss and invoke different flow control mechanisms. TCP and its variants provide reliable connections by retransmitting the lost packets. However, the resulting latency is in general too large for real-time and streaming media applications. For this reason, most streaming applications use UDP protocol with an unreliable packet delivery. However, by discarding corrupted packets, UDP does not distinguish between packet losses due to congestion and corruption. Alternatively, UDP-Lite applies partial checksum to some parts of a packet (i.e., packet payload) and reduces packet loss rate. [46] It is explicitly designed for certain applications, multimedia for example, which can detect and even recover from certain level of errors. CUDP conducts a precise error detection and recovery through error location information from link-layer. [47]

Note that the transport layer can only adapt effectively if it can observe the network layer and link layer conditions, propagate the information to the application layer, and in the meantime, identify and accommodate the application layer's need.

5.6.3 Network Layer and Link Layer Transmission Adaptation

The application characteristics, such as QoS requirement and packet priority, could be used in coordinating the network layer and link layer. In particular, the persistence level of the link layer ARQ mechanism should adapt to each application's latency and reliability requirements, while the link layer scheduler allocates radio resources to various packet flows based on their QoS priorities. The adaptation, however, requires the link layer and network layer to distinguish different packet flows, which in general can be achieved by an explicit indication of the QoS requirement associated with each packet flow. [48] Note that in some systems, the transport layer and link layer both conduct error recovery by using FEC coding and retransmissions. The balance between both schemes is important for the optimal usage of the overall communication resources. [49] Meanwhile, the network could operate efficiently by using the link layer and physical layer information, such as rate fluctuation and error condition, to distribute channel resources.

5.6.4 Network and Channel Condition Estimation and Report

The adaptation relies on each layer's ability to estimate current and even future network and channel conditions. The receiving entity evaluates current condition to invoke reception mechanism accordingly, while the sending entity uses current and future condition to adjust transmission flow. A condition report based on receiver feedback is normally more accurate than estimations at the transmitter.

Within a protocol stack, the link layer must detect its present status, including link availability, congestion, and error conditions, and signal it to upper layers for appropriate adaptation. In Zheng and Boyce, [50] the receiving link layer formats the location of channel errors in a meaningful manner, either implicitly or explicitly, so that the upper layers can identify and use it to detect and recover channel errors. Network layer and transport layer must propagate signals of the current conditions issued by lower layer(s) and themselves to upper layer(s). A proper form of the information exchange across multiple layers is crucial to the effectiveness of the adaptation.

5.6.5 Proxy Server

To allow efficient packet delivery through heterogeneous networks, a proxy server or gateway is placed between different networks. It provides seamless connection between the application server and the end users, regardless of their underlying network behavior. Using mobile streaming video as an example, a proxy server at the edge of the wireless network can virtually separate transmission path to server-to-proxy (e.g., wireline) and proxy-to-mobile-user (e.g., wireless). It can transcode media signal to a format suitable for low rate wireless transmission and limited mobile display, [51] add channel coding or perform retransmissions to maintain reliable transmissions, [52], [53] and prefetch portions of media signals to allow continuous playback during adverse channel conditions. [54] In addition, a proxy server can monitor network conditions in different paths and feedback them to the application server for appropriate adaptation. [55]

In general, in building an efficient wireless network, we strive to create a series of protocol layers that communicate, interact, and thus yield continuously improved applications and services. Next, we will highlight some of the innovative processes to improve the performance of streaming video over wireless network in terms of adaptation and cross layer design.

[33]A multilayered approach to mobile networking, Stanford University Project Report.

[34]Tong, L., Zhao, Q., and Mergen, G., Multipacket reception in random access wireless networks: from signal processing to optimal medium access control, IEEE Communications Magazine, 39 (11), 108–112, 2001.

[35]Girod, B. and F rber, N., Wireless Video, in Compressed Video Over Network, Reibman, A.and Sun, M.T., Eds., Marcel Dekker, New York, 2000.

[36]Girod, B. et al., Advances in channel adaptive video streaming, Proc. IEEE International Conference on Image Processing (ICIP 2002), Rochester, September 2002.

[37]Liu H. and Zarki, M.E., Adaptive source rate control for real-time wireless video transmission, Mobile Networks Appl., 3, 49–60, 1998.

[38]Aramvith, S., Pao, I.-M., and Sun, M.-T., A rate-control scheme for video transport over wireless channels, IEEE Trans. Circuits Syst. Video Technol., 11, 569–580, 2001.

[39]Hsu, C.Y., Ortega, A., and Khansari, M., Rate control for robust video transmission over burst error wireless channels, IEEE J. Selected Areas Commun., 17 (5), 756–773, 1999.

[40]Cen, S., Cosman, P.C., and Voelker, G.M., End-to-end differentiation of congestion and wireless losses, SPIE Multimedia Computing and Networking, (MMCN2002), San Jose, CA, January 18–25, 2002.

[41]Samaraweera, N.K.G., Noncongestion packet loss detection for TCP error recovery using wireless links, IEEE Proc. Commun., 146 (4), 1999.

[42]Biaz, S. and Vaidya, N., Discriminating congestion losses from wireless losses using interarrival timers at the receiver, Technical report 98-014, Computer Science Department, Texas A&M University, June 1998.

[43]Bikram, S. et al., Improving performance of TCP over wireless networks, Technical report 96-014, Texas A&M University, 1996.

[44]Balakrishnan, H. et al., A comparison of mechanisms for improving TCP performance over wireless links, IEEE/ACM Trans. Networking, 5(6), 756–759, 1997.

[45]Goff, T. et al., Freeze-TCP: A true end-to-end TCP enhancement mechanism for mobile environments, in Proc. of IEEE Infocom 2000.

[46]Larzon, L., Degermark, M., and Pink, S., Efficient Use of Wireless Bandwidth for Multimedia Applications, MoMuc '99, San Diego, November 1999, pp. 187–193.

[47]Zheng, H. and Boyce, J., An improved UDP protocol for video transmission over Internet-to-wireless networks, IEEE Trans. Multimedia, 3 (3), 356–364, 2001.

[48]Advice to link designers on link Automatic Repeat reQuest (ARQ), Internet Draft, March 2002, draft-ietf-pilc-link-arq-issues-04.txt.

[49]Chockalingam, A. and Bao, G., Performance of TCP/RLP protocol stack on correlated fading DS-CDMA wireless links, IEEE Trans. Vehicular Technology, 49, 28–33, 2000.

[50]Zheng, H. and Boyce, J., An improved UDP protocol for video transmission over Internet-to-wireless networks, IEEE Trans. Multimedia, 3 (3), 356–364, 2001.

[51]de los Reyes, G., Reibman, A.R., and Chang, S.F., Error resilient transcoding for video over wireless channels, IEEE J. Selected Areas Commun., 18 (6), 1063–1074, 2000.

[52]Vass, J. et al., Mobile video communications in wireless environments, IEEE International Workshop on Multimedia Signal Processing, Copenhagen, Denmark, Sept. 13–15, 1999.

[53]Pei, Y. and Modestino, J.W., Robust packet video transmission over heterogenous wired-to-wireless IP networks using ALF together with edge proxies, Proc. European Wireless 2002, Feb. 25–28, Florence, Italy.

[54]Fitzek, F.H.P, and Reisslein, M., A prefetching protocol for continuous media streaming in wireless environments, IEEE J. Selected Areas Commun., 19 (10), 2015–2028, 2001.

[55]Yu, F. et al., QoS adaptive proxy caching for multimedia streaming over the Internet, 1st IEEE Pacific Rim Conference on Multimedia (IEEE PCM 2000), December 2000, Australia.




Wireless Internet Handbook. Technologies, Standards and Applications
Wireless Internet Handbook: Technologies, Standards, and Applications (Internet and Communications)
ISBN: 0849315026
EAN: 2147483647
Year: 2003
Pages: 239

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net