Relaxed Concurrent Data Structure Semantics for Scalable Data Processing

Digitization, collaborative applications and data analysis is changing our computing models which are becoming more data oriented, with data handling being at the center of computations.

The respective software systems operate under stringent constraints for response times, throughput, scalability, and availability.

Traditional computing methods for sharing data in collaborative applications and for handling data in data analysis are reaching their limits.So far we used high level and rigor semantics for describing the behavior of algorithms and data structures.

Thisis making scaling difficult as data volume and sharing increases and traditional deterministic designs are reaching their limits. Semantic relaxation is a promising way to overcome these computational limits.

Relaxed semantics contain deliberate and controled deviations from the desired semantics and may imply a reduction in accuracy of the system.

By relaxing the semantics, a bigger design space, that allows weaker synchronization and more useful parallelism, is unveiled.

Investigating new data structure designs, capable of trading semantics for achieving better performance in a monotonic way, is a major challenge in the area.In order to address this problem this research project proposes to study and develop new abstractions, data structures  and algorithms operating under elastic relaxed semantics.