Utopia

There has been significant progress made in the area of wireless networking to realise the vision of ensuring seamless connectivity and mobility. Most current networking approaches, mainly wireless technologies, have been incremental improvements of the packet radio technology introduced in late 70s. Each component of the network is so tightly coupled that any failure or unavailability of a single component might render the entire network dysfunctional. In 2002, one of the grand challenges put forth for computer science and engineering was building evolving complex systems based on many simple unreliable components. One such endeavour is through amorphous networking which involves realising coherent behaviour from vast number of such components interacting asynchronously and locally in a networking system.

UTOPIA aims at developing technologies and systems for Autonomous In-Network Processing with simple cellular-level components. The overarching goal is to empower networks with the ability to self-organise and self-manage in such ways that they could perform sophisticated application specific tasks utilising the spare resources within the networks themselves. This would allow any communication networks to have the ability to operate in an autonomous fashion, thus overcoming scalability issues as the complexity of networks increase. The wireless technologies of special interest for purposes other than traditional public access networks or enterprise networks are wireless mesh and wireless sensor networks.

Short Term

A main short-term outcome is a platform for EANP akin to the Globus platform in the context of GRID computing. Our platform will be constructed in a way to allow its implementation on vastly different platforms such as wireless sensor nodes and high-end computers. Thus, the core technology is brutally slim-lined and highly extensible. We construct the specifications for the platform base on two sub-projects within Utopia: TITAN and (ARC application). The two projects share the demand for in-network processing. However, the platforms and goals of the processing differ significantly. The challenge is to construct a generic set of tools and aprocessing environment suitable for both applications.

Long Term

The long-term goal of the project is to develop a scientifically well-founded theoretical study of the possibilities and limitations inherent in massive distributed systems coupled with the realisation of technologies and techniques derived from the theoretical studies. We aim at studying such complex systems on a broad front, determining a suitable framework for the empowerment of such networks through Embedded Autonomous Network Processing (EANP). The notion of EANP is the usage of spare resources within the network itself, for the network itself and its applications. We aim at constructing a set of generic building blocks and techniques that can be used to tailor network behaviour for a given scenario in massively complex networks infrastructure similar to the role autonomous protocols play in traditional communication network infrastructure to enable generic functions. Our main focus is not the protocols themselves but rather the algorithmic underpinnings for system configuration and maintenance.