Carousel: Low-Latency Transaction Processing for Globally-Distributed Data

Abstract

The trend towards global applications and services has created an increasing demand for transaction processing on globally-distributed data. Many database systems, such as Spanner and CockroachDB, support distributed transactions but require a large number of wide-area network roundtrips to commit each transaction and ensure the transaction's state is durably replicated across multiple datacenters. This can significantly increase transaction completion time, resulting in developers replacing database-level transactions with their own error-prone application-level solutions. This paper introduces Carousel, a distributed database system that provides low-latency transaction processing for multi-partition globally-distributed transactions. Carousel shortens transaction processing time by reducing the number of sequential wide-area network round trips required to commit a transaction and replicate its results while maintaining serializability. This is possible in part by using information about a transaction's potential write set to enable transaction processing, including any necessary remote read operations, to overlap with 2PC and state replication. Carousel further reduces transaction completion time by introducing a consensus protocol that can perform state replication in parallel with 2PC. For a multi-partition 2-round Fixed-set Interactive (2FI) transaction, Carousel requires at most two wide-area network roundtrips to commit the transaction when there are no failures, and only one round trip in the common case if local replicas are available.