A Quorum-Based Replication Framework for Distributed Software Transactional Memory

Distributed software transactional memory (D-STM) promises to alleviate difficulties with lock-based (distributed) synchronization and object performance bottlenecks in distributed systems. Past single copy data-flow (SC) D-STM proposals keep only one writable copy of each object in the system and are not fault-tolerant in the presence of network node/link failures in large-scale distributed systems. In this paper, we propose a quorum-based replication (QR) D-STM model, which provides provable fault-tolerant property without incurring high communication overhead compared with SC model. QR model operates on an overlay tree constructed on a metric-space failure-prone network where communication cost between nodes forms a metric. QR model stores object replicas in a tree quorum system, where two quorums intersect if one of them is a write quorum, and ensures the consistency among replicas at commit-time. The communication cost of an operation in QR model is proportional to the communication cost from the requesting node to its closest read or write quorum. In the presence of node failures, QR model exhibits high availability and degrades gracefully when the number of failed nodes increases, with reasonable higher communication cost.