Scalability for Dummies

Part 1 Clone

• Load balancer - evenly distributes load onto group/cluster of application servers

Part 2 Database

• Scalable database solution
  • DBA, master-slave replication (read from slaves, write to master)
    • expensive, time consuming
  • demoralize right from the beginning and include no more joins in any database query!
    • switch to a better and easier scale NoSQL database (e.g. MongoDB, CouchDB)

Part 3 Cache

• In-memory
• 2 Patterns of caching data
  • Cached database queries
    • Hash the result of queries
    • Issue: expiration, perturbation
  • Cached objects
    • directly store the data array/complete instance of class
    • make async processing possible
• Objects to cache
  • user sessions (never use the database!)
  • fully rendered blog articles
  • activity streams
  • user <-> friend relationships

• Part 4 Asynchronism

• wait for a while

  • doing the time-consuming work in advance and serving the finished work with a low request time
    • massive framework, pre-rendered and locally stored
  • unforeseen tasks
    • send the signal when job is done