HardInfrastructure45 min

Design a Distributed Cache

Design a distributed caching system like Memcached or Redis that provides low-latency data access across a cluster.

AmazonGoogleMetaNetflix

Functional Requirements

Sub-millisecond read/write latency
Horizontal scaling: add nodes to increase capacity
High availability: no single point of failure
Support TTL-based expiration
Handle cache invalidation and consistency

Steps (0/6)

Step 1: Clarify Requirements

What data types? (strings, hashes, lists, sorted sets) What's the expected data size? (~1TB total cache). Expected throughput? (1M ops/sec). Read/write ratio? (80:20). Consistency requirements? (eventual consistency OK for most use cases).

Key Points

Sub-millisecond latency for both reads and writes
~1TB total cache across cluster
~1M operations/second
Eventual consistency is acceptable

1 / 6