PPAA 2014 - Workshop on Parallel Programming for Analytics Applications

Analytics applications are scaling rapidly in terms of the size and variety of data analyzed, the complexity of models explored and tested, and the number of analytics professionals or data scientists supported concurrently. Homeland security, consumer behavior modeling, IT infrastructure security and resiliency, and fraud detection and prevention are examples of application areas where the scaling is stressing the computational capabilities of current systems. At the same time hardware systems are embracing new technologies like on-chip and off-chip accelerators, vector extensions to the instruction sets, and solid state disks. New programming methodologies and run-times to support them are emerging to facilitate the development of the new analytics applications, and to leverage the emerging systems. This workshop provides a forum for the applications community, run-time and development environment community and systems community to exchange the outlook for progress in each of these areas and exchange ideas on how to cross leverage the progress. Topics of interest include, but are not limited to:
Parallelism in Social Media and other large Graph Analytics
o Graph databases
o Computational models and programming languages for large graph applications
o Application Frameworks for large graph applications
Large Scale Interactive Analytics.
o Tools for interactive analytics
o DSL extensions and query interfaces for interaction with large graphs
o System and operating system support for interactive analytics. Scheduling, storage hierarchy configurations/layout, network bandwidth and topology
System/Hardware support for Big Data analytics
o Distributed File Systems and NoSQL databases
o The parallel programming models, languages and frameworks for Big Data
Concurrency in large tabular data analytics
o Columnar databases, large data warehouses, data cubes and OLAP engines
o In memory analysis for real-time queries on large data
Exploitation of GPUs, FPGAs and on-chip vector processing units for analytics applications
Parallel I/O, efficient exploitation of the memory hierarchy, particularly solid state disks
System management issues for attaining the desired levels of reliability and performance for the above application/system combinations
Parallel algorithms for large graphs, other big data and other analytics applications, for attaining the performance goals
Analytics applications and solutions in homeland security, marketing and customer loyalty, healthcare and other industri