RPOs of Seconds: Is Near-Zero Data Loss Realistic?

The recovery point objective, or RPO, defines how much data an organization can afford to lose when something goes wrong. Traditional backup strategies commonly accepted RPOs of hours, sometimes a full day. For many workloads today, that is no longer acceptable. Organizations increasingly ask whether they can lose only seconds of data, or none at all. The answer is that near-zero data loss is realistic, but it involves trade-offs that deserve careful consideration.

What Determines RPO

RPO depends on how frequently data is captured and copied to a protected location. Nightly backups produce RPOs of up to a day. Hourly snapshots produce RPOs of up to an hour. Continuous replication, capturing every write as it occurs, can reduce RPO to seconds.

Synchronous Replication

Synchronous replication writes data to two locations simultaneously and confirms the write only when both have received it. This achieves an RPO of zero, since no confirmed write can be lost. However, it requires very low latency between sites, which typically limits distance to a metropolitan area. Applications also experience slightly higher write latency.

Asynchronous Continuous Replication

Asynchronous continuous replication sends changes to the recovery site immediately after they are written locally, without waiting for confirmation. This allows much greater distances between sites with little impact on application performance. RPOs are typically measured in seconds, depending on network bandwidth and change rates. For most organizations, this approach offers an excellent balance of protection, performance and distance.

Network Requirements

Continuous replication requires sufficient bandwidth to keep up with the rate of data change. If bandwidth is inadequate, replication falls behind and RPO grows. Careful sizing based on actual change rates is essential.

Cost and Complexity

Achieving very low RPOs requires investment in replication software, network capacity and recovery infrastructure. Organizations should reserve the most demanding protection for systems where data loss carries the greatest cost.

Beyond Disasters

Near-zero RPO also matters for ransomware recovery. Continuous protection with journaling allows recovery to a point just before an attack, minimizing lost data.

Matching RPO to Business Need

Classifying applications by the business impact of data loss helps assign appropriate RPOs and avoids overspending on less critical systems.

Monitoring is equally important. Replication platforms should report actual RPO continuously, alerting administrators when lag grows beyond acceptable limits. Without this visibility, an organization may believe it is protected to the second when replication has quietly fallen minutes or hours behind due to a network issue or an unexpected surge in data changes. Dashboards that show replication lag alongside change rates help teams spot trends before they become problems and justify network upgrades with evidence.

The Bottom Line

RPOs measured in seconds are entirely realistic with continuous replication. Synchronous replication can achieve zero data loss over short distances, while asynchronous replication delivers seconds over longer ones. Network capacity and cost are the main constraints. Matching protection levels to business impact ensures resources are spent wisely. For critical systems, near-zero data loss has moved from aspiration to achievable standard.