ACIA Transactions

With online applications requiring high availability, ACIA (Atomicity, Consistency, Isolation, Availability) is now becoming the new standard for transaction support, instead of ACID (Atomicity, Consistency, Isolation, Durability). Given ACIA transactions, more diverse requirements can be flexibly supported for applications through the specification of consistency levels, isolation levels and fault tolerance levels. Clarifying the ACIA properties enables the exploitation of techniques used for ACID transactions, as well as bringing about new challenges for research. The following presents the specification of ACIA properties.

  • Atomicity. The effects of either all or none of the operations in a transaction are reflected in the database, with the user knowing which of the two results is.
  • Consistency. Each successful transaction can commit only legal results, preserving the consistency of the database. Legal results comply with rules and consistency levels specified for the database.
  • Isolation. Operations within a transaction must be isolated from other transactions running concurrently. How much transactions can be isolated from other transactions is defined as the isolation levels, which are guaranteed by different concurrency control mechanisms.
  • Availability. Once a transaction has committed its results, the system must guarantee that these results are reflected in the database, whose data can be accessed by any client connected to the system.

Position Paper:

  • Yuqing ZHU, Jianxun Liu, Mengying Guo, Wenlong Ma, Guolei Yi, Yungang Bao. ACIA, not ACID: Conditions, Properties and Challenges. CoRR abs/1701.07512, 2017.

Related Articles:

  • Implementation Proposals
    • Yuqing ZHU, Jianxun Liu, Wenlong Ma, Mengying Guo, Yungang Bao. Transaction Support over Redis: An Overview. CoRR abs/1702.00311, 2017.
  • Atomic Commit
    • Yuqing ZHU, Philip S. Yu, Guolei Yi, Mengying Guo, Wenlong Ma, Jianxun Liu, Yungang Bao. Logless One-Phase Commit Made Possible for Highly-Available Datastores. Accepted to Distributed and Parallel Databases (DAPD) (pdf).
    • Yuqing ZHU, Philip S. Yu, Guolei Yi, Wenlong Ma, Mengying Guo, Jianxun Liu. To Vote Before Decide: A Logless One-Phase Commit Protocol for Highly-Available Datastores. CoRR abs/1701.02408, 2017.
    • Yuqing ZHU. Non-Blocking One-Phase Commit Made Possible for Distributed Transactions over Replicated Data (poster). 2015 IEEE International Conference on Big Data, IEEE BigData, 2015.
  • Concurrency Control
    • Yuqing ZHU, Yilei Wang. Improving Transaction Processing Performance By Consensus Reduction. 2015 IEEE International Conference on Big Data, IEEE BigData, 2015.
    • Yuqing ZHU, Yilei Wang. SHAFT: Serializable, Highly Available and Fault Tolerant Concurrency Control For Large-Scale Datastores. The 21st IEEE International Conference on Parallel and Distributed Systems, ICPADS, 2015.
    • Yuqing ZHU. SHAFT: Serializable, Highly Available and Fault Tolerant Concurrency Control in the Cloud. Technical Report (ICT, CAS), 2013. (link)
  • Replica Control
    • Yuqing ZHU, Philip S. Yu, Jianmin WANG. RECODS: Replica Consistency-On-Demand Store (demo).IEEE International Conference on Data Engineering, ICDE, 2013, Pages 1360-1363.
    • Yuqing ZHU, Jianmin WANG, Philip S. Yu. Malleable Flow for Time-Bounded Replica Consistency Control. 10th USENIX Symposium on Operating System Design and Implementation (OSDI '12 poster). Hollywood, CA, USA, October, 2012. (proposal, poster)
    • Yuqing ZHU, Philip S. Yu, Jianmin Wang. Latency Bounding by Trading off Consistency in NoSQL Store: A Staging and Stepwise Approach. CoRR abs/1212.1046, 2012.