Abhijeet Mohapatra

A technique to incrementally update compressed relations in column stores. A compressed attribute is updated in time that is logarithmic in the number of the distinct values of that attribute. Batch updates can be performed in time that is constant per updated tuple. No technique can perform faster updates in an asymptotic sense.

Architecture that is an index mechanism which optimizes complex queries that result from enforcing fine grained access control. The architecture addresses the problem of efficient query evaluation in the presence of fine grained access control.

A method for compressing a database relation by leveraging an optimal ordering of tuples and composition of compression operators.

Proposed and implemented a mechanism to optimize fine-grained access control in database queries. For a variety of authorization policies, an order of magnitude speed-up was observed on TPC-H benchmark. (USPTO Publication#: 20120330925)

Proposed and implemented an extensible framework to compress database relations by composing compression techniques. The compression achieved on syntenthic as well as real datasets using our framework was an order of magnitude larger than state-of-art compressors such as WinZip and xMill. (USPTO Publication#: 20120150877)

(a) Proposed a novel architecture which decouples the sequencial modules in a traditional game loop.
(b) Proposed a new algorithm for collision-free navigation in crowd simulations.

Developed a formal model to answer continuous queries on complex event streams in the temporal logic framerwork of Lustre.