Today, big data systems deal with volume, velocity, variety and veracity of the application data which may be deployed on dedicated single high performance systems (such as Netezza), dedicated commodity based clusters or shared architectures such as clouds. We witness an explosive growth in the complexity, diversity, number of deployments and capabilities of big data processing systems such as Map-Reduce, Hbase, Hive, Cassandra, Big Table, Hyracks, Dryad, Pregel and Mongo DB. The big data system may use new operating system designs, advanced data processing algorithms, parallelization of application, high performance computing architectures such as GPUs etc. and clusters to improve the performance. Looking at the volume of data to mine, and complex architectures, one may need to monitor, analyze, identify or predict bottlenecks to optimize the system and improve its performance. The workshop on performance analysis of big data systems (PABS) aims at providing a platform for scientific researchers, academicians and practitioners to discuss techniques, models, benchmarks, tools and experiences while dealing with performance issues in big data systems. The primary objective is to discuss performance bottlenecks and improvements during big data analysis using different paradigms, architectures and technologies such as Map-Reduce, Hbase, MPP, Big Table, NOSQL, graph based models and any other new upcoming paradigms. We propose to use this platform as an opportunity to discuss systems, architectures, tools, and optimization algorithms that are parallel in nature and hence make use of advancements to improve the system performance. This workshop shall focus on the performance challenges imposed by big data systems and on the different state-of-the-art solutions proposed to overcome these challenges.