Veeam Home | Support | Downloads
Best Practices for Veeam Backup & Replication 7.0 for VMware

Veeam Management Pack Help Center  > Best Practices for Deployment & Configuration >  Resource Planning and Optimization > Planning for Repositories > Understanding the Impact of IOPS on Backup Performance

Understanding the Impact of IOPS on Backup Performance

Previous page Next page Print this Topic

Table of contents

As discussed in earlier sections, Veeam Backup & Replication offers two primary modes for storing backups to disk, Forward and Reverse incremental. Due to the differences in how these modes write backups to disk they have very different storage I/O profiles. Below we will discuss the impact that IOPS has on each of these modes.

Fisrt, let's define the IOPS and what they mean from a Veeam Backup & Replication perspective. IOPS stands for Input/Output Operations per Second and is a standard measurement used by storage vendors to define the performance of disk and storage solutions. Consider that there are many different factors that can play into this calculation such as I/O size, cache, queue depth, read/write mix, etc. For sizing IOPS with Veeam Backup & Replication it is recommended to assume an average I/O size of 128K with a low queue depth.

Another critical factor is Sequential vs Random IOPS. The performance of many Veeam Backup & Replication functions, such as Reverse Incremental, Synthetic Full, SureBackup and Instant Restore are most impact by the ability of the storage array to deliver Random IOPS. The most important factor for random IOPS are the physical constraints of the disk, things like head movement time and rotational latency, as well as any overhead from RAID or other resiliency features. In general, the faster a disk spins, the higher it’s IOPS measurement is.

The table below lists typical IOPS performance for common disk used in datacenters.

Device

IOPS

7,200 RPM SATA

~75-100 IOPs

10,000 RPM SATA/SAS

~125-150 IOPs

15,000 RPM SAS

~175-210 IOPs

As an example of how RAID settings can impact the available storage capacity as well as random IOPS and throughput let’s look at a simple example of an array with twelve 3TB SATA disk in various RAID configurations using a 256K stripe size.

RAID Level

Approximate Usable Capacity

Approximate Random IOPS (128K)

Approximate Random I/O Throughput

0

32TB

600

75MB/s

10

16TB

450

55MB/s

5

29TB

300

37MB/s

6

26TB

225

28MB/s

The above calculations are based on industry standard algorithms for calculating RAID overhead; however, various architectures can have slightly different impact. You should consult the storage vendor for more detailed information for calculating IOPS for a specific platform. Caching can have a significant impact on performance as well, and storage systems that offer full write-back caching can in many cases mitigate some of the impact of RAID overhead. The numbers above do not take caching into account, since the impact varies from vendor to vendor.

Using this information, you can estimate the number of IOPS required to reach your required performance for the various modes described next.

Table of contents