Introduction
Data backup is an essential part of any organization’s data protection strategy. However, as data volumes continue to grow, it becomes increasingly challenging to back up and restore data within an acceptable timeframe. This article will examine the factors that influence backup and restore performance, provide benchmarks on what speeds are typically achievable, and offer guidance on assessing if your backup system is fast enough for your needs.
Key factors influencing backup and restore performance
Several key factors affect how quickly you can back up and restore data:
Storage system performance
The throughput and IOPS capabilities of your backup storage system set limits on how fast backup and restore operations can run. Faster disk drives, flash storage, parallelism, and compression can all help maximize performance.
Network bandwidth
The available network bandwidth between your production environment and the backup storage impacts transfer speeds. Faster LANs and WANs enable quicker backup and restore.
Backup software capabilities
The capabilities of the backup software itself impact performance. Features like multithreading, parallel streams, and incremental backups can significantly accelerate backups.
Hardware resources
The CPU, memory, and I/O resources allocated to backup infrastructure and software affect how fast they can operate. Underpowered backup servers can bottleneck performance.
Type of backup
Full, incremental, differential, and synthetic backups have different performance profiles. Incremental and synthetic backups are faster than full backups.
Backup and restore speed benchmarks
As a general guideline, here are some benchmarks for backup and restore speeds:
- Disk to disk backup over LAN – Up to 500 MB/s
- Disk to disk backup over WAN – Up to 100 MB/s
- Disk to tape backup – Up to 400 MB/s
- Virtual machine backup – Up to 300 MB/s
- Bare metal restore from disk – Up to 600 MB/s
- File restore from disk – Up to 100,000 files per hour
Speeds vary based on the factors outlined earlier. Faster storage, networks, software, and resources can double, triple, or increase speeds even more.
Assessing if backup performance is adequate
To assess if your current backup system meets your restore objectives, focus on these areas:
-
Recovery point objective (RPO) – How much data loss can you tolerate in the event of a failure? Your backups must complete within the RPO window.
-
Recovery time objective (RTO) – How long can you afford for systems to be down before restores are complete? The RTO determines if restore speeds are sufficient.
-
Backup window – Do your backups consistently complete within the allocated backup window? If not, you may need faster backups.
-
Restore times – Are your actual restore times meeting RTO expectations? Test restores of key data sets and systems.
-
Growth – Are backup speeds maintaining pace as data volumes increase over time? Model expected growth.
-
SLAs – Are backup speeds allowing you to remain within SLA commitments for RPO, RTO, and service availability?
Fallback options like physical tape shipment may be needed if online restores cannot meet RTOs. Ongoing performance monitoring provides insight into when upgrades may be required.
Conclusion
Matching backup and restore performance to business requirements is critical. By understanding key speed factors, establishing realistic benchmarks, and continuously assessing if performance is meeting SLAs, an organization can ensure their backup infrastructure performance is fast enough for their needs, both now and into the future.
