Adaptive Transmission of multimedia data

Introduction
 
The multicast transmission of real time multimedia data is an important component of many current and future emerging Internet applications, like videoconference, distance learning and video distribution. The heterogeneous nature of the Internet makes the multicast transmission of real time multimedia data a challenge. Different receivers of the same multicast stream may have different processing capabilities, different loss tolerance and different bandwidth available in the paths leading to them.
 
The heterogeneous network environment that Internet provides to real time applications as well as the lack of end-to-end QoS (Quality of Service) guarantees, many times forces applications to embody adaptation schemes in order to work efficiently. In addition, any application that transmits data over the Internet should have a friendly behaviour towards the other flows that coexist in today's Internet and especially towards the TCP flows that comprise the majority of flows. A TCP friendly flow consumes no more bandwidth than a TCP connection, which is traversing the same path with that flow.
 
Usage of adaptation mechanisms
 
The implementation of adaptation mechanisms in the applications is often criticized. The main arguments that rise against it, are that the technologies that are used today for the implementation of the core networks provide capabilities to support QoS; as a result the network should offer to the applications QoS guarantees. This is generally true but there is a big problem about it: Today's Internet is divided into thousands of different administration domains. The QoS strategies that are implemented on each one are certainly different (for example QoS based on DiffServ Concept, QoS based on IntServ Concept, QoS based on IPv6 infrastructure), and in many cases no QoS strategy is implemented at all. So the multimedia data flows that have to traverse many of these different domains in order to reach the end user don't have a sufficient QoS support. Another idea widely supported among network administrators, is that the cost of exhaustive monitoring of the network as well as the upgrade of the links that constrain the entire network domain (bottlenecks and critical links) cost less than the deployment of QoS schemes (Research, testing and personnel training).
 
Many researchers urge that due to the use of new technologies for the implementation of the networks, which offer QoS guarantees, adaptive real time applications will not be used in the future. We believe that this is not true and adaptive real time applications will be used in the future for the following reasons: (1) Users may not always want to pay the extra cost for a service with specific QoS guarantees, when they have the capability to access a service with good adaptive behavior, (2) some networks may never be able to provide specific QoS guarantees to the users, (3) even if the Internet eventually supports reservation mechanisms or differentiated services, it is more likely to be on per-class than per-flow basis. Thus, flows are still expected to perform congestion control within their own class. (4) With the use of the differential services network model in the future, networks will support services with QoS guarantees together with best effort services and adaptive services.
 
Methods for adaptive transmission
 
The subject of adaptive transmission of multimedia data over networks has engaged researchers all over the world. During the design and the implementation of an adaptive application special attention must be paid to the following critical modules: (1) the module, which is responsible for the transmission of the multimedia data, (2) the module, which is responsible for monitoring the network conditions and determines the change to the network conditions, (3) the module, which is responsible for the adaptation of the multimedia data to the network changes, (4) the module, which is responsible for handling the transmission errors during the transmission of the multimedia data.
 
Multicast transmission of real time multimedia data is an important component of many current and future emerging Internet applications, like videoconference, distance learning and video distribution. The heterogeneous nature of the Internet makes the multicast transmission of real time multimedia data a challenge. Different clients of the same multicast multimedia data may have different processing capabilities, different loss tolerance and different bandwidth available in the paths leading to them. Should the server let the client with the least capacity dictate the adaptation? Is it fair the server ignores such a client? General speaking the server must treat the group of clients with fairness. The methods proposed for the multicast transmission of time sensitive data in the Internet can be generally divided in three main categories, depending on the number of multicast streams used:
 
  • The Server uses a single multicast stream for all Clients. This results to the most effective use of the network resources, but on the other hand the fairness problem arises.
  • Simulcast: The Server transmits versions of the same multimedia encoded in varying degrees of quality. This results to the creation of a small number of multicast streams with different rates, responsible for a range of Clients with similar capabilities. The different streams carry the same multimedia information but in each one the multimedia is encoded with different bit rates, and even different formats. So each Client joins in the stream that carries the multimedia quality, in terms of bit rate, that it is capable of receiving. The main disadvantage in this case is that the same multimedia information is replicated over the network but recent research has shown that under some condition simulcast behaves better than transmission of layered encoded.
  • The Server uses layered encoded technique, which is multimedia that can be reconstructed from a number of discrete data streams and transmits each layer into different multicast stream. The multimedia is divided in to one basic stream and more additional streams. The basic stream provides the basic quality and the quality improves with each layer added. The Clients subscribe to one or more multicast streams depending on the available bandwidth into the network path to the Server.
 
Our Work
 
Our team has proposed various adaptation mechanism that have been presented in know journals and conferences:
 
Adaptive unicast transmission
"Streaming multimedia data with adaptive QoS characteristics", Ch. Bouras, A. Gkamas, Protocols for Multimedia Systems, 2000, Cracow, Poland, October 22-25 2000, pp. 129-139.
 
"Performance of Adaptive Multimedia Transmission: The case of Unicast Technique", C. Bouras, A. Gkamas, Fifth International Network Conference (INC 2005), Doryssa Bay Resort, Samos Island, Greece, 5th - 7th July 2005 pp. 121 - 128.
 
Adaptive one stream multicast transmission
"Multimedia Transmission with Adaptive QoS based on Real Time Protocols", Ch. Bouras, A. Gkamas, International Journal of Communications Systems, Wiley InterScience, 2003; 16, pp. 225-248.
 
"A Mechanism for Multicast Multimedia Data With Adaptive QoS Characteristics", Ch. Bouras, A. Gkamas, Protocols for Multimedia Systems 2001, October 17-19, 2001, Enschede, The Netherlands, pp. 74-88.
 
"Performance of Adaptive Multimedia Transmission: The case of "One Multicast Stream" Techniques", Ch. Bouras, A. Gkamas, The International Conference on Information Networking (ICOIN) 2004, Busan, Korea, Feb. 18 - 20 2004.
 
Adaptive simulcast transmission
"Architecture and Performance Evaluation for Redundant Multicast Transmission Supporting Adaptive QoS", Ch. Bouras, A. Gkamas, An. Karaliotas, K. Stamos, Multimedia Tools and Applications, Kluwer Academic Publishers, 2005; 25, pp. 85-110.
 
"Architecture and Performance Evaluation for Redundant Multicast Transmission Supporting Adaptive QoS", Ch. Bouras, A. Gkamas, An. Karaliotas, K. Stamos, 9th International Conference on Software, Telecommunications and Computer Networks (SoftCom 2001) Split, Dubrovnik (Croatia), Ancona, Bari (Italy), October 09-12 2001, volume II, pp. 585-592.
 
"An Architecture for Redundant Multicast Transmission Supporting Adaptive QoS", Ch. Bouras, A. Gkamas, An. Karaliotas, K. Stamos, 7th International Workshop on Multimedia Systems, Capri, Italy, November 7-9 2001, pp. 133-142.
 
"Behaviour Investigation using Simulation for Redundant Multicast Transmission Supporting Adaptive QoS", Ch. Bouras, A. Gkamas, An. Karaliotas, K. Stamos, IEEE International Conference on Networks (ICON 2002), Singapore, August 27 - 30 2002, pp. 112-117.
 
"SRAMT - S: A hybrid sender and receiver - based adaptation scheme for TCP friendly multicast transmission using simulcast approach", Ch. Bouras, A. Gkamas, The 2002 IFIP Workshop on Internet Technologies, Applications, and Societal Impact (WITASI - 02), Wroclaw, Poland, October 10 - 11 2002, pp. 105-122.
 
Adaptive layered encoded transmissions:
"SRAMT: A hybrid sender and receiver-based adaptation scheme for TCP friendly multicast transmission", Ch. Bouras, A. Gkamas, Computer Networks Journal, Elsevier Science, Vol. 47, No. 4, March 2005, pp. 551 - 575.
 
"SRAMT - LE: A hybrid sender and receiver - based adaptation scheme for TCP friendly multicast transmission using layered encoding", Ch. Bouras, A. Gkamas, Communication Networks and Distributed Systems Modelling and Simulation Conference (CNDS'03), pp. 71 - 76, January 19 - 23, 2003 Orlando, Florida, pp. 71-76.
 
"Comparing performance of SRAMT-LE vs. other layered encoding schemes regarding TCP friendliness", Ch. Bouras, A. Gkamas, 6th IEEE International Conference on High Speed Networks and Multimedia Communications HSNMC'03 Estoril, Portugal, July 23 - 25 2003, pp. 493-502.
 
Our team
 
The members of our team are:
 
Dr Christos Bouras (Associate Professor) [home page]
Apostolos Gkamas (Ph.D, R&D Computer Engineer) [home page]