Objectives : The overall objective is to coordinate partners’ research activities in areas related to reliable and efficient communication in self-organized networks: documenting and sharing knowledge with respect to existing approaches, identifying new challenges and, most importantly, collaborating on various related research topics, sharing software and performing joint experimentation activities.
Recent advances in wireless communications enable a radical new communication paradigm: self-organised information and communication systems. in this new networking environment, the users’ mobile devices are the network and they must co-operatively provide the functionalities usually provided by the network infrastructure (e.g. routers, switches, and servers). Networks that exploit the self-organization principles are namely the Mobile Ad-hoc Networks (MANETs), Wireless Mesh, Sensor, and Peer-to-Peer networks. More specifically the following the areas will be addressed:
· Cross Layer Design and Optimization. Wireless networks present many challenges that are not present in traditional wired networks. The point-to-point channel capacity may be time-varying and subject to fading; in fact, the channel may be broadcast in nature rather than point to point. Nodes may be mobile resulting in a constantly changing network topology. Moreover, node mobility and the consequent reliance on batteries in order to remain operational, imposes restrictions on power expenditure; hence energy-aware designs and lifetime elongation become of paramount importance. Wireless self-organized networks pose additional challenges; the nature of these networks dictates that they operate with minimal or no dependence on centralized authority, in a distributed manner and relying on local information in order to achieve reliable and efficient network-wide operation. The challenges above render the traditional layered approach to network design questionable. It turns out that in order to achieve the required performance objectives it is necessary to take into account concurrently issues that belong on several layers of the standard network model, an approach that is referred to a “Cross Layer Design”. Due to its importance, a lot of research effort has been dedicated the last years in issues related to Cross layer design. While significant progress has been achieved so far in terms of understanding the fundamental issues involved, a lot of work is still needed in order to bring the design closer to practice, especially concerning self-organized networks.
- Development of Cooperative Networks. in self-organized networks, it often turns out that nodes decide to join the network in order to achieve their own interest, rather than a well-defined global objective. Such is the case, for example, when an ad-hoc network is formed consisting of PCs that belong to a number of individuals with no ties to a particular organization, or in overlay P2P networks. in such cases, design methodologies that are based on optimizing global objectives under the assumption that each node will cooperate fully are ineffective. One must take into account the fact that nodes will act selfishly and will cooperate provided that their own self-interests are satisfied, or at least, are not harmed. The development on networks in such an environment where each node has its own objectives poses many challenges and fits naturally within the game-theoretic framework that has been developed over the years by economists.
- Applications of Network Coding to Self-organized Network. Network coding is a radically new approach to network routing. Traditionally each node in a network is viewed as a device which takes packets of information arriving at its input and delivering to one or more of its output. in this respect information is treated similarly to fluid flow. However, information can also be operated upon, for example two packets can be added together if each is considered as an element of an appropriated field. in general, two or more packets from the inputs of a node may be mapped (translated) to a new packet and then transmitted to the node output(s). A receiving node observes all the received packets and may be able to recover the original packets provided it knows the mappings that took place during the packet’s sojourn in the network. It has been shown that multicasting throughput can be improved if nodes have the capability of performing these mappings. The same holds in a network with multiple unicast connections although the exact performance improvement is not known at present. However, it turns out that Network Coding has additional features that make it a very promising approach in the design of self-organized networks. It can be implemented in a fully distributed manner, it can be robust to topology changes and can render polynomially solvable a number of problems that are NP-hard with conventional approaches. A large amount or research has been dedicated the past few years to addressing fundamental problems related to Network Coding, however there are still unresolved issues. Moreover, a lot of work is still needed in order to make the proposed methodologies practical.