Large-Scale Distributed Systems (LSDS) group has 3 accepted SIGMOD papers

The Large-Scale Distributed Systems group has 3 accepted papers at this year's SIGMOD conference, the premier venue for data management research.

The Large-Scale Distributed Systems (LSDS) group (http://lsds.doc.ic.ac.uk) has three accepted papers at this year's ACM International Conference on the Management of Data (SIGMOD). The annual ACM SIGMOD/PODS Conference is a leading international forum for database researchers, practitioners, developers, and users to explore cutting-edge ideas and results, and to exchange techniques, tools, and experiences. SIGMOD'16 (http://sigmod2016.org) takes place from June 26 to July 1, 2016 in San Francisco, USA.

The LSDS group, led by Dr. Peter Pietzuch (http://www.doc.ic.ac.uk) in the Department of Computing, has the goal to support the design and implementation of tomorrow's large-scale distributed systems. It investigates new abstractions and infrastructures for building scalable, robust and secure data-intensive applications. The group currently consists of eight post-docs and seven PhD students.

The accepted papers by the LSDS group cover a range of topics related to big data processing:

* "SABER: Window-Based Hybrid Stream Processing for Heterogeneous Architectures"

Alexandros Koliousis, Matthias Weidlich, Raul Castro Fernandez, Paolo Costa, Alexander L. Wolf, and Peter Pietzuch (link: http://lsds.doc.ic.ac.uk/content/saber-window-based-hybrid-stream-processing-heterogeneous-architectures )

This paper describes SABER, a new architecture for a parallel streaming engine that can take advantage of heterogeneous processors such as multi-core CPUs and GPUs. By combining the resources of all parallel processors in a server, SABER can achieve substantially higher performance for many streaming workloads.

* "AT-GIS: Highly Parallel Spatial Query Processing with Associative Transducers" 

Peter Ogden, David Thomas, and Peter Pietzuch (link: http://lsds.doc.ic.ac.uk/content/gis-highly-parallel-spatial-query-processing-associative-transducers)

This paper describes AT-GIS, a new approach for the parallel processing of geospatial queries over big data on multi-core CPUs. The idea behind AT-GIS is to express geo-spatial queries using "associative transducers", which can be parallelised with little overhead.

* "THEMIS: Fairness in Federated Stream Processing under Overload"

Evangelia Kalyvianaki, Marco Fiscato, Theodoros Salonidis, and Peter Pietzuch (link: http://lsds.doc.ic.ac.uk/content/themis-fairness-federated-stream-processing-under-overload)

This paper investigates the problem of fairness in federated stream processing deployments, which span several data centres. It describes a new query-agnostic fairness metric that quantifies the value of processed data, and proposes an efficient load shedder based on this metric.

Congratulations to all of the authors in the LSDS group!