Controller-Based Replication

Businesses know that controller-based replication is a time-tested solution for disaster recovery. But what few people understand is the different types of replication and how it meets their needs. Three basic types of controller-based replication exist, synchronous replication, asynchronous replication, and point-in-time replication.

  • Synchronous replication ensures that each write to disk is immediately copied to the remote disaster recovery site. This ensures the best possible RTO and RPO but at the highest cost point, both in terms of solution cost and performance impact. Because servers must wait for synchronous replication to complete before the next task can be performed, synchronous replication is limited to short distance, mission-critical, expensive storage environments.
  • Asynchronous replication introduces a small delay between the time the storage is written to disk at the primary site and at the remote disaster recovery site. The server is allowed to perform at full speed at a slightly degraded RTO and RPO than synchronous replication and a significantly lower cost. As such, it provides the best balance between recovery objectives and cost-performance.
  • Point-in-time replication is a variant of asynchronous replication that creates a “remote snapshot” capability. It has a much worse RTO and RPO because the snapshot loses consistency with the current data at the primary site. It is useful for providing the last-known-working-copy scenario.

The StorStac product family provides a powerful set of asynchronous replication tools, StorAR, that is fully integrated with StorStac management platform. Key features include:

  • Powerfully integrated with the entire StorStac product family
  • Distance-less replication over existing IP infrastructure
  • Works with qualified WAN provisioning solutions

Key applications include:

  • Disaster recovery
  • Data distribution
  • Centralized backup
  • Backup window management
  • Application testing

Definitions:

Recovery Time Objective (RTO) — Period of time in which systems must be up by, a RTO of 5 hours means systems should be operational within 5 hours.

Recovery Point Objective (RPO) — Point in time in which data must be recovered to. For a higher rate of data churn, a lower RPO should be used. An RPO of 5 minutes means you can acceptably recreate 5 minutes of data.

©2009 Intransa, Inc. All Rights Reserved.