The world economy is increasingly evolving into a knowledge based one. The roles of information and content source, provider and consumer are blurring. The computing paradigm is shifting from distributed to decentralized. The whole internet is turning into an intra-planetary repository of information and content. The primary focus of my research is information systems that can scale up to such internet-scale dimensions.

Large-scale systems often either by chance or design exhibit self-organizational properties. I am interested in objective understanding of self-organization using tools from cybernetics, for example, Markov analysis of such systems, and leveraging this understanding to orchestrate self-organization in these scalable systems. My approach is to look into realistic problems (sometimes pertaining to existing systems), abstracting them to a level bereft of artefacts in order to understand the fundamental dynamics of the systems in as general applicability as possible, and porting such understanding in realizing new and better algorithmic and system designs.

Various aspects of scalability:

Performance metrics: Per participant/usage cost, System's cumulative cost, Collateral cost, Guarantees (precision/recall/coverage)
Feasibility: System formation (emergence), Decentralized boot-strapping, Heterogeneity
Sustainability: Fault-tolerance (static resilience), Dynamic equilibrium (operational cost)

SANDS Research group

I lead the Self-* and Algorithmic aspects of Networked Distributed Systems (SANDS) research group.