Skip to Main content Skip to Navigation
Reports

Persisting the AntidoteDB Cache: Design and Implementation of a Cache for a CRDT Datastore

Abstract : Many services, today, rely on Geo-replicated databases. Geo-replication improves performance by moving a copy of the data closer to its usage site. High availability is achieved by maintaining copies of this data in several locations. Performance is gained by distributing the data and allowing multiple requests to be served at once. But, replicating data can lead to an inconsistent global state of the database when updates compete with each other.In this work, we study how a cache is designed and implemented, for a database that prevents state inconsistencies by using CRDTs. Further, we study how this cache can be persisted into a checkpoint store and measure the performance of our design with several benchmarks. The implementation of the system is based on AntidoteDB. An additional library is implemented to realise the discussed design.
Complete list of metadata

https://hal.inria.fr/hal-03654003
Contributor : Ayush Pandey Connect in order to contact the contributor
Submitted on : Friday, April 29, 2022 - 1:16:05 PM
Last modification on : Wednesday, June 8, 2022 - 12:50:06 PM

File

RR-9470.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-03654003, version 2

Citation

Ayush Pandey, Annette Bieniusa, Marc Shapiro. Persisting the AntidoteDB Cache: Design and Implementation of a Cache for a CRDT Datastore. [Research Report] RR-9470, TU Kaiserslautern; LIP6, Sorbonne Université. 2022. ⟨hal-03654003v2⟩

Share

Metrics

Record views

19

Files downloads

8