View consistency for optimistic replication
Ashvin Goel, Calton Pu, Gerald J. Popek
Abstract
Ashvin Goel, Calton Pu, Gerald J. Popek
Abstract
Optimistically replicated systems provide highly available data even when communication between data replicas is unreliable or unavailable. The high availability comes at the cost of allowing inconsistent accesses, since users can read and write old copies of data. Session guarantees have been used to reduce such inconsistencies. They preserve most of the availability benefits of optimistic systems. We generalize session guarantees to apply to persistent as well as distributed entities. We implement these guarantees, called view consistency, on Ficus an optimistically replicated file system. Our implementation enforces consistency on a per-file basis and does not require changes to individual applications. View consistency is enforced by clients accessing the data and thus requires minimal changes to the replicated data servers. We show that view consistency allows access to available and high performing data replicas and can be implemented efficiently. Experimental results show that the consistency overhead for clients ranges from 1% to 8% of application runtime for the benchmarks studied in the prototype system. The benefits of the system are an improvement in access times due to better replica selection and improved consistency guarantees over a purely optimistic system.
OpenAlex reports 14 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Optimistically replicated systems provide highly available data even when communication between data replicas is unreliable or unavailable. The high availability comes at the cost of allowing inconsistent accesses, since users can read and write old copies of data. Session guarantees have been used to reduce such inconsistencies. They preserve most of the availability benefits of optimistic systems. We generalize session guarantees to apply to persistent as well as distributed entities. We implement these guarantees, called view consistency, on Ficus an optimistically replicated file system. Our implementation enforces consistency on a per-file basis and does not require changes to individual applications. View consistency is enforced by clients accessing the data and thus requires minimal changes to the replicated data servers. We show that view consistency allows access to available and high performing data replicas and can be implemented efficiently. Experimental results show that the consistency overhead for clients ranges from 1% to 8% of application runtime for the benchmarks studied in the prototype system. The benefits of the system are an improvement in access times due to better replica selection and improved consistency guarantees over a purely optimistic system.
Key concepts: Computer science, Replica, Consistency (knowledge bases), Eventual consistency, Consistency model, Data consistency, Replication (statistics), Session (web analytics)