Explore a case study on building CloudKit, Apple's cloud storage system for structured data, using FoundationDB and the FDB Record Layer. Delve into key features that have significantly benefited CloudKit, including FoundationDB's multi-key ACID transactions for advanced secondary indexing and transactional full-text search, the Record Layer's version indexes for improved cross-device synchronization scalability, and FoundationDB's reliability for simplified operations and failure recovery. Discover how these technologies enable richer APIs, greater scalability, and higher availability across billions of databases. Learn about record stores, sharding, load balancing, stateless compute, streaming queries, resource management, high concurrency sync, and text indexing in this comprehensive exploration of CloudKit's architecture and implementation.
Using FoundationDB and the FDB Record Layer to Build CloudKit
Overview
Syllabus
Intro
Thousands of Apps, Billions of Databases
FDB, Record Layer, and Record Stores
Record Stores and Sharding
Balancing the Load
Record Layer for Stateless Compute
Streaming Queries
Limits, Throttling, and Resource Management
CloudKit and Indexing
High Concurrency Sync
Record Sync - The Old Way
Record Sync - Everything's Better with Versionstamp
Record Sync With Versionstamp
Text Indexes
FoundationDB Robustness in Practice
Summary
Taught by
Linux Foundation
