Self-organization methodologies for services placement in future mobile communication networks

Ubiquitous access to information and services anywhere, anytime and anyhow at low cost is one of the essential features of future mobile communication networks, which will interconnect a heterogeneity of various systems and be much more dynamic and flexible in terms of changes in access technology, topology, services, etc. In such networks, there will be a need to match resources supply with application demands continuously as these demands are expected to fluctuate over time. This makes an adequate service placement in such networks of a major importance. Handling this issue is the main purpose of our research, which will be carried out in the scope of the International Graduate School on Mobile Communications at Ilmenau University of Technology (www.gs-mobicom.de).

The paper proposes a novel distributed and self-organized service placement mechanism that enables minimizing the cost, retaining the Quality of Service (QoS) and balancing the load. Our proposal is called Mobile Agent-assisted Ant Colony Optimization (MA-ACO). The basic idea lies in enhancing the ACO technique with mechanisms from the MA-field, so that the speed, accuracy, and self-organization capabilities are improved. This has been achieved by means of ants carrying a set of MAs as a piece of food. As a replica of a service is created on a node, this node sends ants to put MAs on the 1-hop neighbors capable of hosting the offered service. When the serving cost increases, QoS degrades, etc., ants are sent towards the MAs to collect information about the current status of the network. Based on this information, the hosting node can take an accurate decision. Our approach can cope with dynamic networks, where network topology, load, clients' distribution, etc. often change. Moreover, it has self-healing capabilities since it is able to remove congestion situation, balance the load, etc. As MA-ACO works in a distributed manner, it does not require global knowledge of the network. Our proposal overcomes existing approaches with respect to self-organization capabilities, speed and accuracy.