Upgrade to Pro — share decks privately, control downloads, hide ads and more …

FoundationDB

 FoundationDB

Quick tour of FoundationDB distributed fault-tolerant ACID database.

_amirouche_

August 31, 2018
Tweet

More Decks by _amirouche_

Other Decks in Programming

Transcript

  1. • data expert system = database • space vs time

    • read vs write • indexing • structure vs query language • operations
  2. History & Ecosystem - MySQL / PostgreSQL / Oracle /

    MSSQL / IBM DB2 / sqlite - Big Data / NoSQL - HBase / Cassandra (CQL) / Riak / MongoDB / CouchDB / DynamoDB / REDIS / ElasticSearch - dbm, bsddb, Kyoto Cabinet, LMDB, LevelDB, RocksDB, WiredTiger - Google Spanner, CockroachDB, TiDB & FoundationDB
  3. Consistent A read sees all previously completed writes. Availability Reads

    and writes always succeed. Partition Guaranteed properties are maintained even when network failures prevent some machines from communicating with others. https://apple.github.io/foundationdb/cap-theorem.html
  4. Features - Scalable - ACID transactions - Fault tolerance -

    Replicated Storage - Ordered Key-Value API - Watches - Atomic Operations - OLTP / OLAP https://apple.github.io/foundationdb/features.htm l
  5. Anti-features - Data models - Query languages - Analytic frameworks

    - Disconnected operation - Long-running read/write transactions https://apple.github.io/foundationdb/anti-features.html
  6. FDB vs PostgreSQL - Similar guarantees - Directory instead of

    tables - Indices are built by layers / applications - Watches instead notify / triggers - Layers instead extensions - Scalability / Distributed Fault Tolerance
  7. FDB vs ElasticSearch - Stronger guarantees - Key-Value instead of

    Document - Indices are built by layers / applications - No map / reduce
  8. FDB vs MongoDB 3.4 - Stronger guarantees - Key-Value instead

    of Document - Indices are built by layers / applications - No map / reduce