NewSQL: concepts, realities and when you should embrace it
summary
Claim: Distributed SQL gives relational SQL and ACID consistency while scaling horizontally. Concrete examples: Google Spanner, CockroachDB, YugabyteDB; architecture uses ranges with automatic splitting, replication (typical replication factor of three), and consensus (Paxos or Raft). Walkmate scenario: a celebrity tweet drives >1M logins causing connection contention on a single Postgres RDS; Distributed SQL removes manual sharding and supports multi-region locality and data residency by pinning rows, but costs more hardware, requires careful partition keys, and risks replication storms or regional data loss if pinned incorrectly.
AI-written summary · about
chapters (5)
- 0:00Start
- 1:54The History of Databases
- 12:06Read and Write Scenario in a Distributed Sql Database
- 13:14Read Write Scenario
- 24:04Should You Embrace Distributed Sql
related talks
29mModern Database Architecture: From SQL to NoSQL & Distributed SQL
12mNewSQL Leveraging PostgreSQL
02h 42mSQL (The Return): Demystifying Preconceptions and Using the DB Effectively
01h 11mThe official ten-year retrospective of NewSQL databases
47mAnatomy of a Distributed SQL Database - YugabyteDB
31mSQL to NoSQL to NewSQL and the rise of polyglot persistence Paul Dix
31mHarness the Power of a Single Database for Microservices
24mTech Roundtable: Distributed SQL Around the World
