The Differences Between Backup and Replication
A-to-Zerto Glossary of Terms
Related Topics
What Are Backup and Replication?
Both these terms are defined in more details in their respective articles, but at a high level here is what they are:
- Backup involves making a copy or copies of data and storing them offsite in case the original is lost or damaged.
- Replication is the act of copying data and then moving data between a company’s sites, whether those be datacenters, colocation facilities, public, or private clouds.
The Main Differences Between Backup and Replication
Choosing the technology that is best suited for your business can be challenging. Fortunately, we have summarized in the diagram below the main differences between backup and replication to explain some of the key distinctions.
Backup
Replication
DEFINITION
Involves making a copy or copies of data at specific points in time.
DEFINITION
Replication is the act of copying and then moving data between a company's sites. It is typically measured in Recovery Time Objective (RTO) and Recovery Point Objective (RPO).
PURPOSE
Backup focuses on compliance and granular recovery, such as long-term archival of business records.
PURPOSE
Replication and recovery focus on Disaster Recovery - quick and easy resumption of operations after an outage or corruption. Minimizing the Recovery Time Objective (RTO) is key.
USES
Backup is typically used for everything in the enterprise, from production servers to desktops.
USES
Replication is often used for mission-critical
applications that must always be up and running.
HOW IT WORKS
Backup typically relies on snapshots which are copies of the data set taken at a pre-determined point-in-time.
HOW IT WORKS
Replication can be synchronous, asynchronous or near-synchronous and may use continuous data protection to enable users to access historic data.
REQUIREMENTS
Backup requires a tape library (usually VTL doing disk-to-disk backup) and some place to store archived tapes.
REQUIREMENTS
Replication requires investment in another
infrastructure in order to enable recovery and continued business operations.
BOTTOM LINE
Backup is a relatively inexpensive way to avoid complete data loss. Valuable for compliance. Does not ensure continuity of operations.
BOTTOM LINE
Replication is focused on ensuring that business applications and processes are always available, even after an outage.
4 Reasons Why Backup is Not Replication
Besides the differences noted above, there are specific reasons that make clear backup is definitely not the same as replication.
- Service Levels. Backups typically occur once per day and at night, which means that the potential data loss could be days or more. When protecting the applications and data that matter to your business, this amount of data loss is unacceptable. Restoring from a backup, especially a tape backup, can take days; from disk it might be slightly faster – a few hours.
- Application Impact. Backups rely on snapshot-based technology. The reason they are taken so infrequently is because this type of technology drains resources on the server. It is possible to take more frequent copies, but this comes at the expense of server resources and-user productivity is significantly impacted.
- Reverse Protection. Once applications and data have been made available at a target site, protection must be extended to include the new data that users are creating. A backup solution will not start taking backups and ship them back to the production site. Replication technologies will replicate back to the source site, ensuring the application is still protected both during and after an outage.
- Retention. Backups are normally stored for a very long time for compliance and audit purposes. Depending on how often they occur, the recovery granularity can be hours, days, or more. Technologies that use continuous data protection (CDP) offer extremely granular recovery points, often separated by mere seconds. This gives your enterprise several points in time to recover to, just in case the last point in time is corrupted.
Frequently Asked Questions
Do I need both backup and disaster recovery?
Yes, it is recommended to run backup and DR solutions in parallel as part of a multi-faceted data protection strategy. Use replication for DR to dramatically limit data loss and downtime during a disruption or ransomware attack—especially important for mission-critical workloads. Many enterprises then use a backup solution for long-term retention, compliance, and archiving, particularly for workloads where longer RPOs and RTOs are tolerable; these backups should typically be made off-hours to avoid performance impact from snapshots.
Which should I use for instant restores of files, folders, or VMs?
Replication solutions with a granular journal will typically be better suited for routine, day-to-day restores (sometimes called “operational recovery”). The last good copy of a file or VM state will be only seconds in the past with replication software rather than 12 or 24+ hours old with backup. Although not likely a disaster per se, a DR solution that can instantly restore files, folders, and VMs directly back to their original location will often be much faster than restoring from a periodic snapshotting tool.
Can I use synchronous replication instead of backup and DR?
No, synchronous replication solutions are not an adequate replacement, although in some situations they can be complementary. They are particularly suited to geographically close locations where latency is not a concern. However, near-synchronous replication with a granular journal ensures easy recoverability even after corrupt data or malware incidents—unlike with synchronous replication, bad data is not immediately propagated to the target site and overwriting previously clean data.
How often should I backup or replicate?
Backup frequency can be highly dependent on business needs, SLAs, and the achievable performance with the backup infrastructure. In general, most enterprises backup daily, usually late at night to avoid disrupting production during working hours. Top-tier replication solutions do not require scheduling, don’t require windows of operation, and are always-on. Their continuous replication ensures 24x7 coverage so that no application goes unprotected.
Will replication impact production performance?
A software-only, hypervisor-based replication solution will not impact production performance. The near-synchronous replication that makes up continuous data protection (CDP) only transmits changed blocks, so it is more akin to a trickle of data rather than a deluge. Additionally, enterprise-class replication solutions capture writes before they’re written to storage so there is no interruption in the data path. Replication can act as a low overhead background process that keeps business protected without disrupting the performance of servers and applications.
Zerto for Backup and Replication
Backups are relatively inexpensive ways to maintain some level of protection. However, today’s digital consumers demand an always-on level of service that backups simply cannot provide. Zerto is changing the way companies think about protecting their data, providing a single solution for all your data protection needs.
Zerto offers a newer, more advanced backup technology—called journaling technology—that lets you replicate data continuously and in real time, recording all your data every few seconds and giving you the lowest RPOs in the industry.
Zerto Solution Overview
Other Resources
LATEST FROM ZERTO SEE ALL
Data Replication
Learn more about data replication.
Backup & Recovery
Learn more about backup and recovery.
Backup and Disaster Recovery: Are The Lines Blurring?
While backup is not replication, how different is it from disaster recovery?
What is Zerto?
Learn about Zerto and how it can help you solve your data protection and recovery challenges.