The Hyflow project is developing concurrency control abstractions, protocols, and mechanisms for current and emerging multicore architectures, cluster systems, and geographically distributed systems. A particular focus is to understand what concurrency control abstractions promote high programmability for these architectures, and how to support those programming abstractions with high performance, scalability, and dependability. A closely related focus is to build open-source experimental systems that embody the techniques, as well as incorporate them into existing open-source infrastructures (e.g., Infinispan).
Ongoing efforts are exploring high performance fault-tolerant transactional memory for cluster systems, high performance software transactional memory for multicore architectures, scalable transactional data structures, and scalable transactional protocols for geographically distributed systems.
Recent Selected Publications
- "Archie: A Speculative Replicated Transactional System", S. Hirve, R. Palmieri and B. Ravindran, ACM/IFIP/USENIX 15th International Middleware Conference (MIDDLEWARE 2014), December 8-12, 2014, Bordeaux, France
- "Brief Announcement: On Breaching the Wall of Impossibility Results on Disjoint-Access Parallel STM", S. Peluso, R. Palmieri, P. Romano, B. Ravindran and F. Quaglia, 28th International Symposium on Distributed Computing (DISC 2014), October 12-15, 2014, Austin, Texas