Beyond Eventual Consistency

on YouTube ↗46m

" summary

Starbucks rebuilds a unified commerce platform on Scala Akka Kafka Cassandra and an AWS pass and evaluates global stores: Aurora, Cosmos, Spanner, Cockroach and FaunaDB. FaunaDB (Calvin) POC writes 100 million transactions with no observed consistency failures despite node failures, and Starbucks deploys the new system in Japan adding 2 million new rewards customers. The concrete tradeoffs: single-region strong serializability versus latency and operational limits.

AI-written summary · about

" chapters (6)

1. 0:00Start
2. 17:27Amazon Aurora
3. 19:57Azure Cosmos
4. 23:50Google Spanner
5. 29:38FaunaDB - things to know
6. 37:23Cassandra

" related talks

• 01h 00mBeyond Eventual ConsistencyJamie Allen · Scala.IO 2017
• 43mEffective Eventual Consistency with Actor Models + Amazon Web ServicesPhilip Laureano · NDC Sydney 2017
• 51mImmutable data stores and CORS for safety, flexibility and profitSidney Shek · Scala Days Berlin 2016
• 40mReturn of the Transaction KingRyan Knight · Scale By the Bay 2017
• 26mBuilding Financial Systems on Eventually Consistent DBsRahul Pilani · Strange Loop 2018
• 01h 01mLost in transaction Strategies to manage consistency in distributed systemsBernd Rücker · JPoint 2019
• 46mLost in transaction? Strategies to deal with (in-)consistency in distributed systemsBernd Rücker · Full Stack Fest 2018
• 32mDiscover the Modern Way to Build Real-Time Analytics Solutions at Any ScaleGolan Nahum, Ilya Gulman · hayaData 2024