How Do I Know How Big a NAS to Buy?

Many of the same considerations that you would give any storage device should be applied to the acquisition of a network-attached storage (NAS) device, and a good capacity plan should be at the top of the list.

The planning process is probably the most difficult but most important task when considering the purchase of additional capacity for any storage device. Fundamentally, capacity planning for the current year is based on the historic growth rate and change rate over time. This is then extrapolated over time to predict future required capacity. Your NAS vendor can help you with native tools to gather the necessary statistics to determine the change rate and track your growth.

Along with factors that will multiply your capacity requirements, such as replication to another device or capacity needed for disaster recovery, remember to take into consideration factors that will decrease your requirements, such as data classification or data profiling exercises. In some cases, this may eliminate up to 70 percent of the stored and backed up data by identifying it as either redundant or non-business data, e.g., MP3 files. Analyzing the data first can reduce the cost of the buy and usually streamlines the data management process.

Outside of the capacity-planning discussion from above, there are several differences when storing data in a NAS device instead of a direct attached storage device that utilizes iSCSI or FCP. It is important to understand and accommodate for the various overheads. The first overhead is the RAID, which varies from 5 percent to 20 percent. From this, there can be an additional 10 percent overhead for the file system for housekeeping activities.

Also, it is important to understand how much data will be locked within the snapshot. A snapshot is essentially used for recording a moment in time. Some vendors simply write duplicate data to a separate area, while others manipulate a pointer at the block level to point to the live data, thus requiring less disk.Understand (1) the data change rate, (2) the duration of time the snapshots will be retained, and (3) the way the snapshot data is stored for your particular NAS vendor in order to predict the future capacity. It is not unusual that this can be about 25 percent or more of the device depending on your change rate.

In brief, assuming you have an active capacity planning process when calculating the size of hardware purchase of a NAS device, consider the overheads that are required for the technology; additional required space for RAID, file system, and snapshots. As seen in the table below, 165 Gbytes of raw disk are required for every 100 Gbytes of usable disk, but obviously your mileage will vary.

Table 1: