Spanner Google’s Globally-Distributed Database 斯潘纳：谷歌的全球分布数据库.pdf

Spanner: Google’s Globally-Distributed Database James C. Corbett, Jeffrey Dean, Michael Epstein, Andrew Fikes, Christopher Frost, JJ Furman, Sanjay Ghemawat, Andrey Gubarev, Christopher Heiser, Peter Hochschild, Wilson Hsieh, Sebastian Kanthak, Eugene Kogan, Hongyi Li, Alexander Lloyd, Sergey Melnik, David Mwaura, David Nagle, Sean Quinlan, Rajesh Rao, Lindsay Rolig, Yasushi Saito, Michal Szymaniak, Christopher Taylor, Ruth Wang, Dale Woodford Google, Inc. Abstract tency over higher availability, as long as they can survive 1 or 2 datacenter failures. Spanner is Google’s scalable, multi-version, globally- distributed, and synchronously-replicated database. It is Spanner’s main focus is managing cross-datacenter the first system to distribute data at global scale and sup- replicated data, but we have also spent a great deal of port externally-consistent distributed transactions. This time in designing and implementing important database paper describes how Spanner is structured, its feature set, features on top of our distributed-systems infrastructure. the rationale underlying various design decisions, and a Even though many projects happily use Bigtable [9], we novel time API that exposes clock uncertainty. This API have also consistently received complaints from users and its implementation are critical to supporting exter- that Bigtable can be difficult to use for some kinds of ap- nal consistency and a variety of powerful features: non- plications: those that have complex, evolving schemas, blocking reads in the past, lock-free read-only transac- or those that want strong consistency in the presence of tions, and atomic schema cha