✅Strong Consistency

NEW

A guarantee that reads return the most recent committed value system-wide, often equivalent to linearizability for single items.

DESCRIPTION

Strong consistency ensures that once a write is acknowledged, all subsequent reads observe it. It simplifies reasoning about correctness and invariants but typically increases latency, coordination overhead, and sensitivity to partitions. Systems achieve strong consistency via consensus, synchronous replication, or strict quorum configurations. It is common in transactional databases and services requiring correctness guarantees like financial ledgers.

ALTERNATIVE DEFINITIONS

Google Cloud Spanner

A strong read is guaranteed to see the effects of all transactions that have committed before the start of the read.

Clarifies temporal guarantee at read start time for global consistency.

Source

Microsoft Azure

Strong consistency offers the highest consistency. Reads are guaranteed to return the most recent committed version of an item.

Ties strong consistency to linearizability semantics in Cosmos DB.

Source

RELATED TERMS

Linearizability Serializability Quorum Consensus Google Spanner Azure Cosmos DB

COMPETING TERMS

Eventual Consistency

RELATIONS

• Broader: Consistency

KNOWLEDGE AREAS

Data Storage & Operations Data Architecture Data Engineering Software Engineering

TARGET AUDIENCE

Data Architect Cloud Architect Data Engineer Software Developer Data Modeler

EXTERNAL SOURCES

• Wikipedia - Consistency model
• Google Cloud - Spanner strong reads documentation
• CockroachDB - Serializable consistency overview

TOP QUESTIONS

When should I require Strong Consistency?

Use it for correctness-critical operations like financial transfers, inventory control, or reference updates where stale reads break invariants.

Category: Requirements

What are the trade-offs of Strong Consistency?

It often increases latency and reduces availability under partitions due to coordination, impacting throughput and tail latencies.

Category: Risks

Is Strong Consistency the same as Linearizability?

For single-key operations, strong consistency typically implies linearizability; multi-key transactions may require serializability guarantees.

Category: Misunderstanding

How is Strong Consistency implemented?

Via consensus protocols (e.g., Raft, Paxos), synchronous replication, strict quorums, or globally synchronized timestamps.

Category: Management

Can I get Strong Consistency across regions?

Yes, but expect higher write/read latencies due to cross-region coordination; systems like Spanner optimize this with TrueTime-like mechanisms.

Category: Opinion

What is a common mistake with Strong Consistency?

Applying it universally. Many read paths can tolerate staleness, where weaker consistency significantly reduces latency and costs.

Category: Mistakes

How do quorums relate to Strong Consistency?

Majority quorums combined with consensus ensure a single, globally agreed sequence of writes observed by all readers.

Category: Concept

How can I validate Strong Consistency?

Use linearizability tests, Jepsen-like fault-injection, and monitor replica divergence and commit timestamps.

Category: Testing

Is Strong Consistency always more expensive?

Generally yes in latency/throughput, but careful design and locality-aware routing can mitigate overheads for common cases.

Category: Opinion