Persistence Basics

-

What does persistence mean

• Data needed by a program to operate sticks around (“Persists”) through restarts/re-deploy
• Important information needed across multiple runs is stored somewhere

-

persistence options

• Run the program; copy & paste the output
• Never restart the program
• Read/write to file
• Use a database

-

Example: Bank Accounts

• Deploying new software shouldn’t reset accounts
• Bugs or crashes shouldn’t impact customer balances
• Multiple applications may need access to account data

-

Separation of Concerns

Your application shouldn’t be responsible for storing data, only knowing where to find it and what to do with it.

Databases live/run for a long time; programs may change daily/hourly

-

Database types

-

Relational Databases

• Mysql, sqlite, Postgresql
• Store data in rows in tables
• Most (but not all) use SQL or similar
• Tables have a fixed schema (set of columns)
• Changing schema is slow and costly

-

Schemaless databases

• eg: MongoDB
• Flexible semantics of entries in the DB
• Still have an implicit schema (code expects certain properties)
• Schema can change on the fly (requires careful management though)

-

Graph Databases

• eg: Neo4J
• Store data as a graph of relationships (eg Facebook’s Social graph)

-

Extra Resources

• Databases on TeachYourselfCS
• Jennifer Widom’s Stanford Database Course