Distributed computing column 54 transactional memory: models and algorithms


This issue’s column consists of a review article by Gokarna Sharma and Costas Busch on models and algorithms for transactional memory (TM), with particular emphasis on scheduling. With the ever-growing popularity of TM, this is a timely topic. The authors cover three main models. First, work on transaction scheduling algorithms for tightly-coupled systems are surveyed. Second, distributed networks systems are considered; the new aspect is how to find the shared objects efficiently and provide consistency of the objects after transactions terminate. Third, related results for non-uniform memory access systems are surveyed, with emphasis on how to provide consistency in a load-balanced way. The article closes with a discussion of future directions. You might want to check out previous coverage of TM in this Column, dating back to 2008. In March of that year, the entire column was devoted to the topic, in the context of multicore systems. The first four Workshops on the Theory of Transactional Memory (WTTM) are reviewed in the December 2009, December 2010, March 2012, and December 2012 issues. The latter issue also covers the awarding of the 2012 Dijkstra Prize for the work by Herlihy, Moss, Shavit and Touitou on TM. Many thanks to Gokarna and Costas for their contribution!

DOI: 10.1145/2636805.2636823

Extracted Key Phrases

11 Figures and Tables

Cite this paper

@article{Welch2014DistributedCC, title={Distributed computing column 54 transactional memory: models and algorithms}, author={Jennifer L. Welch}, journal={SIGACT News}, year={2014}, volume={45}, pages={73} }