Subhajit is a Postdoctoral Researcher working with the Distributed Systems research group at INESC-ID affiliated with University of Lisbon. He is currently working on a project funded by the European Research Consortium to develop theroretical model of consistency and isolation levels of distributed systems under the supervision of Professor Rodrigo Rodriguez.
Subhajit completed his Phd in Computer Science from Louisiana State University in Ausgust 2016. His dissertation was tittled "The performance trade off and SLA awareness in cloud computing and distributed datastores".
Subhajit publishes in top conferences and journals in Computer Science. His papers titled ''OptEx: A Deadline-Aware Cost Optimization Model for Spark'' and ''OptCon: An Adaptable SLA-Aware Consistency Tuning Framework for Quorum-based Stores'' were accepted in the 16th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGrid 2016) 2016.He also worked as a Research Intern for Robert Bosch GmbH, RTC, Bangalore, India in the summer of 2014, with the Text Analytics research team.
Subhajit also has 4 years of experience as software developer at IBM and PWC India, in java, databases, and internet technologies.
Visit Subhajit's Linkedin profile
1. Subhajit Sidhanta, Wojciech Golab, Supratik Mukhopadhyay, Saikat Basu, OptCon: An Adaptable SLA-Aware Consistency Tuning Framework for Quorum-based Stores, 16th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGrid) 2016.
2. Subhajit Sidhanta, Wojciech Golab, Supratik Mukhopadhyay, OptEx: A Deadline-Aware Cost Optimization Model for Spark, 16th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGrid) 2016.
3. Subhajit Sidhanta, Supratik Mukhopadhyay, Managing A Cloud for Multi-agent Systems on Ad-hoc Networks, Work In Progress, 5th IEEE International Conference on Cloud Computing (CLOUD) 2012.
Visit Google Scholar
Because of the exponentially increasing complex nature and intensity of the operations being performed by the modern generation of software applications, traditional hosting environments like workstations, grids, and in some cases, even supercomputers are proving to be more and more insufficient. These circumstances are prompting organisations to look towards moving their operations base from traditional in-house servers and grids to public and private cloud platforms, like Amazon AWS and Microsoft Azure. On the other hand, we are witnessing a similar transformation in the frontier of data storage backends, where the traditional relational database (i.e., RDBMS) solutions are fast making way for distributed storage systems, like Apache Cassandra and Microsoft Azure. RDBMS based systems fail to handle the challenges in storing and querying exponentially increasing workload.
My work aims at: 1) providing organisations with smart tools that estimates the composition of the cloud infrastructure required for hosting the given application workload, and 2) providing automated tuning of the distributed storage systems for meeting the SLA, while maximizing the throughput. I plan to attack the problem using two approaches: 1) develop empirical performance models, and 2) use machine learning making predictions about optimal configuration required in running a given workload.
Subhajit collaborates with academicians and researchers from institutes and industries. He has collaborated with researchers from University of Waterloo, the City University of New York, NEC Labs, and Robert Bosch, India.
A Framework for Modelling Consistency and Isolation Levels:
Consistify: A Framework for Safe and Fair Execution of Concurrent SLA-driven Client Applications on Quorum-based Datastores (under submssion):
OptCon: a Flexible Workload and SLA-Aware Framework for Consistency Tuning (Paper accepted in CCGrid 2016):
OptEx: A Deadline-Aware Cost Optimization Model for Spark (Paper accepted in CCGrid 2016):
YCSB Workload Variations Simulator
View Subhajit's OpenSource contributions in Github