Self-* Aspects of Networked Distributed Systems

At SANDS, we study scalability, security, social and self-organizational aspects of networked distributed systems, hence the name S* Aspects of Networked Distributed Systems, or in brief, SANDS. The SANDS research group, led by Asst. Prof. Anwitaman Datta at the School of Computer Engineering in NTU Singapore includes members coming from the Parallel and Distributed Computing Center (PDCC) and the Center for Advanced Information Systems (CAIS).

We believe that the 21st century knowledge based society will continue to be driven by large-scale networked distributed systems which are self-organizing and adaptive to their environment, and will be strongly coupled to the social bonds and activities it facilitates and serves. Such large-scale networked distributed systems include peer-to-peer systems, online social & collaboration networks as well as the cloud infrastructure. We investigate self-organization and algorithmic issues of these systems and networks and their scalability, resilience, adaptivity, security and performance under dynamic and wide-spectrum of environments. Specifically, our works span the following aspects:

Systems research

Our primary focus is to design, implement and deploy scalable systems. More specifically, we are interested in data management, reliability and security aspects of such systems.

Algorithms

Our systems research is driven by the study and use of distributed and nature inspired algorithms, including randomized, epidemic and swarm ones.

Self-* properties

In order to build scalable and robust decentralized systems, it is essential to understand (or algorithmically design) emergent phenomena like self-organization, stabilization, healing and optimization. We use analytical models and large-scale simulations to identify (or exploit) networked distributed systems' structural properties, time-evolution, causal and cyclic (feedback) relations.

Social networks and informatics

The internet has over the decades changed the way the society interacts and works. The boundaries of the society and the web have turned further amorphous with the advent of Web 2.0 sites and online social networking services. Sociological studies are thus an essential ingredient in understanding modern networked distributed systems. Our work in this realm include social network mining, trust and reputation mechanisms, collaboration communities, information propagation and visualization techniques for semantic social networks.