Title

AWS re:Invent 2023 - Building mission-critical applications with CockroachDB (ARC101)

Summary

Andrew Baptist from Cockroach Labs presents on building mission-critical applications with CockroachDB.
Evolution of Databases: From single-node systems to NoSQL and managed databases, and now to distributed SQL databases like CockroachDB.
CockroachDB offers large-scale, resilient, distributed SQL database capabilities with serializable consistency, ideal for mission-critical applications.
Ideal Customers: Those who cannot afford even short downtimes, as it costs them significantly (e.g., being featured on the Wall Street Journal for an hour of downtime).
Architecture: A distributed SQL layer on top of a distributed key-value store, with Postgres compatibility for easy migration.
Scalability and Resilience: Automatic range splitting for scalability, Raft consensus for resilience, and easy node addition/removal.
Data Locality: Ensures low latency for global applications, with features like regional by row and global tables.
Serializable Isolation: Default isolation level to avoid anomalies, with read committed isolation being introduced for transition purposes.
Follower Reads: Allows slightly stale reads for lower latency and reduced contention.
Deployment Options: Self-hosted, dedicated cloud, or serverless, with over 300 customers across various industries.
Resources: Blog, releases, community Slack, and source code availability.

Distributed SQL Databases: The future of databases lies in distributed systems that can offer both scalability and strong consistency, which is what CockroachDB aims to provide.
Mission-Critical Applications: CockroachDB is designed for applications where downtime is unacceptable, highlighting the importance of database resilience in modern software.
Compatibility and Migration: Emphasis on Postgres compatibility suggests a strategy to ease migration from traditional databases to CockroachDB.
Data Locality and Performance: The discussion on data locality and latency optimization techniques indicates the importance of performance in distributed applications and the challenges in achieving it.
CAP Theorem: CockroachDB's approach to the CAP theorem suggests that while theoretical limitations exist, practical solutions can provide both consistency and high availability.
Adoption and Migration: The mention of over 300 customers and the transition from other systems to CockroachDB indicates a growing trend in the adoption of distributed SQL databases.
Open Source and Proprietary Mix: The availability of source code and a core open-source version, alongside proprietary extensions, reflects a hybrid model that could appeal to a broad range of users and developers.