Databases which, why and usage tips

Databases which, why and usage tips

• 1.
  Databases - Which,Why and Usage Tips Avner Cohen Infrastructure & Architecture @ Fiverr Birdwatcher @ Sewage Pools @avnerner AvnerCohen @ Github
• 2.
  RDBMS Relational DatabaseManagement System
• 3.
  NoSQL Not OnlySQL - but mainly, not RDBMS
• 4.
  Acronyms CAP theorem “.. it is impossible for a distributed computer system to simultaneously provide all three of the following guarantees Consistency, Availability, Partition tolerance “ - Eric Brewer, 1998 ACID “.. a set of properties that guarantee that database transactions are processed reliably - Atomicity - Consistency - Isolation - Durability” BASE “.. an alternative to ACID - Basic Availability, Soft-state, Eventual consistency “
• 5.
  NewSQL Solving scalabilitywithout compromising on ACID HOW? .. CAP Theorem isn’t simply a case of consistency, availability, partition tolerance. Pick two…
• 6.
  Call me Maybe by Kyle Kingsbury • “Breaking distributed systems so you don't have to.” • https://github.com/aphyr/jepsen • http://aphyr.com/tags/jepsen
• 8.
  Now What ?
• 9.
  @ Fiverr
• 10.
  MySQL Properties •Traditional RDBMS • Owns by Oracle and widely used • Open source • Pluggable storage engines (MyISAM, InnoDB, Federated, more) Stuff etc. •Multiple forks (MariaDB, PerconaDB, WebScaleSQL) •Master/Slave setup and pitfalls (Octopus, DBCharmer, MaxScale, Mysql Proxy) •Master/Master - Galera Cluster •Shlomi Noach’s http://code.openark.org/forge/openark-kit, https://github.com/outbrain/propagator •Rails 4 default - strict: false
• 11.
  MongoDB Properties •Document based • BSON (Binary JSON) • Open source • Indices - primary and secondaries • No Joins Stuff etc. •Console - https://github.com/TylerBrock/mongo-hacker •UI - Mongohub (up to date fork - https://github.com/ jeromelebel/MongoHub-Mac) •Eventual consistent, configurable, query level •Gems: Moped, Mongoid, mongo-driver, BSON •TokuMX- https://github.com/Tokutek/mongo •_id can (and should) be overwritten.
• 12.
  PostgreSQL Properties •Traditional RDBMS • (Still) Open Source • Semi structured (JSON) data Stuff etc. • The Rails goto nowadays. • “..Some people use <NoSQL solution> Just for the schemaless feature..”
• 13.
  Redis Properties •REmote DIctionary Server • Data Structure server, single threaded • In memory but also persisted to disk • Master/Slave (Cluster in 4yo beta, but is coming.) • One man show Stuff etc. • Know your date-types (Sorted Sets, Arrays, etc.) • Redis-faina - https://github.com/facebookarchive/redis-faina • BGSAVE forking issues • HyperLogLog - http://antirez.com/news/75 • Do not kill -9 a diskeless master… - https://stripe.com/blog/ game-day-exercises-at-stripe (a.k.a the stripe-redis drama)
• 14.
  Memcached Properties •As basic as it gets Key Value • in-memory • client side sharding • Efficient in large data blocks Stuff etc. • Dalli • Compression • MasterCacheKey • Name Spacing
• 15.
  Couchbase Properties •Document store • Memcached layer out of the box • Views • Operation tooling Stuff etc. • Querying is a PITA • “Thick” clients • LivePerson and Viber
• 16.
  Apache Solr Properties • Search Server • Server on top of Lucene • Fuzzy search • Vibrant and productive community Stuff etc. • Query slave, index master • Extended DisMax • Indexing solutions • Micro service
• 17.
  Wrapping up •Polyglot Persistence - Martin Fowler • Use as a micro services enabler • Learn from Kyle - Embrace Failure
• 18.
  Thank You. @avnerner AvnerCohen @ Github israbirding@gmail.com