Amazon ElastiCache

Amazon ElastiCache is a managed in-memory caching service offering Redis OSS, Valkey (the fork Amazon backs after Redis's license change), and Memcached. It offloads hot reads from databases, accelerates session state, and supports pub/sub and streams — all with sub-millisecond latency.

Engine Choices:

• ElastiCache for Valkey: Amazon's default going forward — open-source continuation of Redis 7.2, fully protocol-compatible with Redis clients. Priced ~20% below Redis OSS on AWS.
• ElastiCache for Redis OSS: The classic option — still supported for existing deployments. Adds replication, Multi-AZ, cluster mode, and persistence.
• ElastiCache Serverless: On-demand variant (Valkey/Redis/Memcached) that auto-scales capacity and charges per-GB-hour plus per-request.
• ElastiCache for Memcached: Simpler, multithreaded cache for basic key-value use cases without persistence or replication.

Key Features:

• Sub-Millisecond Latency: In-memory data paths suitable for caching, leaderboards, rate limiting, and real-time analytics.
• Cluster Mode & Sharding: Horizontal scaling across shards with online resharding; replicas per shard for read scaling and HA.
• Multi-AZ with Automatic Failover: Primary failures promote a replica in seconds.
• Data Tiering: Large node types (r6gd) spill cold data to local NVMe at lower cost per GB.
• Encryption & VPC Isolation: TLS in transit, KMS encryption at rest, subnet groups, and security groups.
• Global Datastore (Redis/Valkey): Cross-region replication for low-latency reads in multiple regions.

Common Use Cases:

• Read-Through / Cache-Aside: Cache hot queries in front of RDS, Aurora, or DynamoDB to reduce load and latency.
• Session Store: Externalize web session state so application servers are stateless.
• Leaderboards & Counters: Sorted sets and atomic counters for gaming and real-time rankings.
• Rate Limiting: Token-bucket counters with TTL for API gateways.
• Pub/Sub & Streams: Lightweight message fan-out for real-time features.