Advances in information technology have resulted in the emergence of new environments, such as mobile ad hoc networks, wireless sensor and actor networks, large-scale Grid and peer-to-peer networks, which in essence are constituted by networked autonomous nodes. Although these environments present enormous potential to facilitate new applications and services, they also pose difficult design challenges--these environments are intrinsically dynamic, unreliable, and large scale. Traditional design approaches that assume that the system is composed of reliable components, and/or that the system is of relatively small scale are not applicable in such environments. In addition, approaches that are based on central and/or explicit control over the system as a whole either introduce a single point of failure or make the system not adaptable, which are not applicable in such environments either. It is therefore critical to explore new design paradigms and approaches that do not suffer from these defects.
The phenomenon of self-organization is pervasive in nature, where biological organisms self-organize a large number of unreliable and dynamically changing components to develop diverse functions. In addition, these biological organisms possess the desirable properties of robustness to failure of individual components, adaptivity to changing conditions, and lack of explicit central coordination. This research seeks inspiration from the study of swarm behavior in nature, such as slime mold, to design and analyze robust, autonomic networking protocols for the aforementioned environments. This project aims to develop autonomic networking protocols based upon bottom-up modeling of simple, interacting units. The advantage to this approach is that one can reduce the dimensionality of the complex system to a small set of primitive functions and parameters governing the simple units that comprise the system. In particular, the investigators seek to model the behavior of slime mold physarum plasmodium to design autonomic networking protocols for wireless sensor and actor networks.
