Speaker: Gokarna Sharma, Assistant Professor in the Department of Computer Science at Kent State University
Title: Scheduling in Distributed Transactional Memory Systems
Abstract: Programming multi-core systems is a difficult andchallenging task. Transactional memory (TM) is a promising new mechanism that aims to replace conventional locking (and hence its many drawbacks) for programming multi-core systems. Using TM, code is split into transactions, blocks of code that appear to execute atomically with respect to one another. Transactions are executed speculatively: synchronization conflicts or failures may cause an executing transaction to abort: its effects are rolled back and the transaction is restarted. In the absence of conflicts or failures, a transaction typically commits, causing its effects to become visible. Several commercial processors provide direct hardware support for TM, including Intel’s Haswell and IBM’s Blue Gene/Q, zEnterprise EC12, and Power8. There are proposals for adapting TM to clusters of GPUs. In this talk, I will focus on the problem of scheduling transactions in distributed systems where transactions residing at nodes of a communication graph operate on shared, mobile objects. A transaction requests the objects it needs, executes once those objects have been assembled, and then possibly forwards those objects to other waiting transactions. I will discuss how the objects’ transfer in the network affects the completion time of all transactions and the total communication cost. Particularly, I will show that there are problem instances for which there is no scheduling algorithm that can simultaneously minimize the completion time and communication cost. I will then discuss distributed TM systems appropriate for rack-scale or cluster-scale networks of nodes and show that (near-)optimal completion time schedules (ignoring the total communication cost) can be obtained. Finally, I will talk about other related issues and possible future work in this direction.
Bio: Dr. Sharma is currently an Assistant Professor in the Department of Computer Science at Kent State University (KSU), Kent, OH, USA. Prior to joining KSU in Fall 2015, he was a postdoctoral researcher in the Computer Science and Engineering division (CSE) at Louisiana State University (LSU) in Baton Rouge, LA, USA, where he also received his Ph.D. in Computer Science in 2014. He received his dual degree European Master of Science in Computer Science from Free University of Bozen-Bolzano (FUB), Italy and Vienna University of Technology, Austria in 2008, and Bachelor degree in Computer Engineering from Tribhuvan University, Nepal, in 2005. He interned as a Summer Consultant at Alcatel-Lucent Bell Labs in Summer 2008. His current research interests include parallel and distributed computing, sensor networks, emerging technologies such as internet of things, and network, graph, and robotic algorithms. Website: http://www.cs.kent.edu/~sharma/