Architecture of a Reliable In-MemorDBMS: the Tarantool Example
summary
Tarantool is used as an example to show how to build a reliable in‑memory DBMS that avoids the common associations of in‑memory systems with cache, non‑persistence, and unreliability. The talk describes architecture and mechanisms (durable storage, replication, crash recovery and operational practices) that make an in‑memory database both fast and persistent for high‑load systems.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 1:29Why in-memory DBs
- 7:55Tarantool evolution
- 13:56Tarantool architecture
- 21:45Persistence and fsync
- 26:04Why replication
- 31:02How replication works
- 36:02Read problems and MVCC
- 41:04Tarantool architecture summary
- 42:08Q&A
related talks
49mPrimary Use Cases for In-Memory DBMS Illustrated by Tarantool
46mTarantool: How to Save a Million Dollars on Your DB
29mTarantool as a Platform for Microservices
38mHow to Forget About Performance Problems? Tarantool as the DBMS
47mStored Procedures in NoSQL DBMSs: The Tarantool Example
56mUsing Tarantool as a Data Virtualization Platform
23mSynchronizing Data from PgSQL to Tarantool
51mReplication in Tarantool: Configuration and Usage
