CompTIA DataSys+ (DS0-001)Business ContinuityMedium

A data engineer is configuring a database system for disaster recovery. The business requires the ability to quickly recover from a total site failure, but they are willing to accept a few hours of data loss if it means significantly reducing infrastructure costs. Which replication type would be most appropriate for this scenario?

  1. ASynchronous Replication
  2. BAsynchronous Replication
  3. CPeer-to-Peer Replication
  4. DTwo-Phase Commit
Show answer & explanation

Correct answer: B. Asynchronous Replication

Asynchronous replication allows the primary database to commit transactions independently of the secondary. This introduces potential for data loss (acceptable 'few hours' RPO) but significantly reduces the performance overhead on the primary and network requirements, thus lowering infrastructure costs compared to synchronous replication.

Why the other options are wrong

  • A. Synchronous replication guarantees zero data loss but requires high network bandwidth and low latency, increasing costs and potentially impacting primary performance. It doesn't align with 'significantly reducing infrastructure costs' while accepting 'few hours of data loss'.
  • C. Peer-to-Peer replication allows all nodes to be primary, increasing complexity and often not the most cost-effective for a simple disaster recovery setup with an acceptable RPO.
  • D. Two-Phase Commit is a protocol for distributed transactions, not a replication type for disaster recovery.

Asynchronous Replication

Asynchronous replication allows the primary database to commit transactions and respond to the application before the data is fully replicated to the secondary database. This offers better performance and lower network requirements but introduces a risk of data loss (non-zero RPO) if the primary fails before replication completes.

  • Primary commits independently of secondary.
  • Lower latency and network requirements than synchronous.
  • Acceptable for non-zero RPO (potential data loss).

Memory trick: Async: cheaper, faster writes, but data loss bites.

More Business Continuity questions