Skip to content

AWS Elasticache

Amazon ElastiCache

  • Elasticache is used to get managed Redis or Memcached.
  • Caches are in-memory databases with really high performance, low latency.
  • Elasticache helps reduce load off of databases for read intensive workloads.
  • Helps make Applications Stateless
  • It is a managed service. AWS takes care of OS maintenance / patching, optimization, setup, configuration, monitoring, failure recovery & backups

Warning

Using Elasticache involves heavy application code changes

Elasticache Solution Architecture

  • Application queries ElastiCache, if not available gets from RDS & store in ElastiCache
  • Data available in cache: Cache Hit, Data not available: Cache Miss (Fetches data from DB & writes to cache)
  • Cache must have an invalidation strategy to make sure only the most current data is used there

User Session Store

  • User logs into any of the application
  • The application writes the session data into ElastiCache
  • The user hits another instance of the application & the instance in the backend retrieves the data & the user stays logged in

Redis vs Memcached

Redis

  • Multi AZ with Auto-Failover
  • Read Replicas to scale reads & have HA
  • Data Durability using AOF persistence
  • Backup & Restore features
  • IAM Authentication is supported

Memcached

  • Multi-node for partitioning of data (Sharding)
  • No high availability (Replication)
  • Non persistent
  • No backup & restore
  • Multi-threaded architecture
  • Only username/password supported

Note

Redis for HA, backup & Memcached when you can afford to loose the data.

ElastiCache Security

  • Elasticache Support IAM Authentication for Redis
  • IAM Policies on Elasticache are only used for AWS API level security

Redis Auth

  • You can set password/token while creating a redis cluster
  • This is an additional layer of security on top of security groups
  • Support SSL in flight encryption

Memcached

  • Support SASL-based authentications

Elasticache Patterns

Lazy Loading

  • All the read data is cached, data can become stale in cache

Write Through

  • Adds or update data in cache when written to a DB (no stale data)

Session Store

  • Store temporary session data in cache (Using TTL)

Info

Lazy Loading: Only read data from DB if there is a Cache Miss

Use Cases

Gaming leaderboards

  • Redis Sorted Sets: Guarantee both uniqueness & element ordering
  • Each time a new element added, it's ranked in real time, then added in the correct order