It can take up to 12 hours or more to process a full backup job and move the entire backup volume to tape or disk. For many companies, that's just too much time. IT departments can use snapshots to capture entire storage environments, or simply track changes to the state of mission-critical systems. Snapshots can establish recovery points in just a fraction of the time needed for a traditional backup and can significantly reduce recovery point objectives (RPOs) by supporting more frequent recovery points. When properly implemented, snapshot data can be used to restore lost files or recover from more substantial data loss. Features like data deduplication are also appearing in snapshots to reduce disk storage requirements, allowing for many more snapshots and longer retention periods. Below are a series of best practices to help you get the most from snapshot technology.
Start by understanding your RPO
It's important to know your RPO and match the snapshot scheme accordingly. For example, if your RPO is one hour, you should be taking snapshots every hour. However, it's important to note that snapshots can cause additional storage traffic each time the snapshot process is executed. Frequent snapshots (e.g., more than one per hour) can possibly cause brief disruptions to other storage service-level agreements (SLAs). So, there is no real benefit to taking more snapshots than called for by the RPO. If snapshots impact storage performance, you'll need to reconsider the RPO or select a different storage target that can accommodate performance requirements.
Balance storage requirements and retention implications
It's important to understand the size of a snapshot, the snapshot frequency and the desired retention. For example, an organization that captures 1 GB snapshots once per hour will need approximately 24 GB per day of snapshot storage. If snapshots need to be retained for up to 60 days, that's a total capacity of 1.5 TB. By comparison, an organization that takes 1 GB snapshots twice per day will need only 2 GB of storage. A retention period of 45 days will lower the total snapshot storage requirements to 90 GB.
Make the most of data deduplication features
Data deduplication technology compresses data by eliminating redundant files, blocks and bits. Effective compression rates of 30-to-1 are routinely quoted, with ratios of 50-to-1 reported when deduplication has been working over time. At a 30-to-1 compression average, a 750 GB drive has the effective storage capacity of 22.5 TB. This allows IT organizations to reduce spending on new storage, and store data for much longer retention periods. If your backup tool does not include native data deduplication, consider implementing deduplication as a separate part of the snapshot scheme -- either deduplicating primary data first, or deduplicating the snapshot once it's stored.
Employ security features to manage recoverability
Traditional backups are usually not recoverable by everyday users, but snapshots can often allow end users to recover the most recent copies of their files without the direct intervention of a storage administrator. While this provides convenience, it also presents a security risk. It's important to prevent users from recovering files that they're not authorized to access.
Integrate snapshots into the broader data protection scheme
Evaluate the importance of your snapshots and determine the level of data protection that is appropriate for your snapshot data. RAID 5 or even RAID 6 protection is usually adequate in the snapshot storage system itself, while some organizations take it further and opt to backup the snapshots with more traditional processes -- but multiple copies of snapshots are rarely necessary or practical. Few storage administrators see the wisdom in backing up snapshots that will likely "age out" of storage in just a few months.
Use heterogeneous snapshot software
Many storage systems ship with some type of snapshot software. This may greatly simplify taking or storing snapshots on that particular storage system, but does not guarantee interoperability with other storage systems in your environment. Instead, opt for a heterogeneous snapshot tool that can see and work with all of the storage platforms that you need to protect with snapshots. This normally isn't an issue when using storage systems from a single vendor because the tool that ships will probably support all storage from that same vendor. With a mixed storage environment, however, you may need to opt for a third-party tool.
Store snapshots separately
One common oversight when taking snapshots is that many administrators wind up storing the snapshots on the same storage platform as the original data. This may ease performance concerns across the LAN or SAN, but the inherent risk to your data cannot be ignored. Save the snapshots to a different storage platform in the data center, or move the snapshots to a remote storage platform across the WAN. If you absolutely must store snapshots on the same array with original data, be sure to consider snapshots in your backup or other data protection scheme.
Be selective about what to snapshot
In today's busy storage environment, it's typically not necessary to back up everything in each process. The same principle is true in snapshots. For example, you may opt to use snapshot tools to protect only key applications such as Oracle databases or Microsoft Exchange records, and then protect more routine data with more traditional backups. Less data means faster snapshots and quicker recoveries.