Synopsis
Understand how to manage data access with precision and speed in distributed databases.
This work explains a novel approach called the Hierarchical Timestamping Algorithm, which uses a hierarchy of data partitions to coordinate transactions without slowing down everyone.
This book introduces a partitioned view of data that helps transactions run more concurrently. It blends ideas from timestamping and locking, showing how a transaction might use different timestamps depending on which data partition it touches. It also compares this method with traditional locking protocols and other timestamping schemes, highlighting tradeoffs in performance and complexity.
- How data is decomposed into partitions and organized into a hierarchy for better concurrency control.
- Rules for classifying transactions based on the partitions they access and how those decisions affect execution order.
- Three protocols that govern reads, writes, and conflicts, and when each is used to optimize throughput.
- The theoretical basis for correctness, including how serializability is maintained in a hierarchical setup.
Ideal for readers of database theory and practitioners seeking deeper insight into advanced concurrency control techniques.
"synopsis" may belong to another edition of this title.
