In February Storage Switzerland held two webinars that used similar techniques to address an age old problem: data protection. Both of these webinars asked primary storage to take a larger role in its own protection. One webinar discussed using snapshots as a big part of the data protection process, the other discussed using copy data. Both of these webinars were very well attended and as a result we’ve been asked “which approach should I use” more than once. In this column I will try to provide some guidance on making that decision.
Snapshot technology has been around for over a decade and is well vetted and understood. Integration with applications like Oracle, MS-SQL and Exchange are readily available. The key for its use as a data protection solution is making sure that the snapshot data is on a second independent system. This means that the storage solution has to be very cost effective, since buying two or more primary storage systems is not an option for most companies.
Watch the on demand webinar "Why 2015 is the Year of Copy Data"
The Snapshot Challenge
Most storage systems, thanks to their use of re-directed snapshots, can maintain thousands of snapshots without impacting performance. There are three challenges in using snapshots as the sole data protection solution. The first is actually finding the data that’s being protected by a snapshot. If all that’s needed is a recovery of the most recent copy of data, this can be done via timestamps. But as those snapshots get older finding that data is increasingly complex. For this reason we suggest backing these snapshots up with traditional backup software so that the data can be indexed and search capabilities can be improved.
The second challenge is moving this snapshotted data to less expensive media over time (deduplicated disk, tape or cloud). Again, a backup solution can deliver one or all of these options to help drive down the cost of storing this data. But in both cases backup remains another, separate process that has to be run.
The third challenge is the presentation of these snapshots to other processes like test-dev-ops, reporting and analytics. With a snapshot-only approach, the presentation of this data to these processes either has to be manually done or done via a series of complex scripts.
Enter Copy Data
Copy data solutions could be the ideal way to address these challenges. Some of these can provide very detailed indexing of snapshot data for improved search, some can provide a “tape-out” functionality and at least one solution can automate, without scripting, the presentation of these snapshot copies to the various tasks described above.
The Types of Copy Data Solutions
There are three types of copy data solutions and it is important to understand the types that are coming to market. Some will allow you to start with the snapshot-driven approach described above and in our webinar, and others will replace it.
The first type essentially splits write traffic and sends data to two storage systems simultaneously, your normal primary storage and your secondary storage. This type is essentially a replacement for the data center’s snapshot technology and should be considered as part of a data protection solution refresh.
The secondary copy is then snapshotted and replicated to be able to provide point-in-time copies and disaster protection. The second type integrates at an application level and then replicates data in near-real time to a secondary storage solution. As with the first type of copy data, this solution essentially replaces current snapshot technology.
The third type leverages the snapshot and replication of the current storage solution but adds the key missing element, the ability to find data and support restores. This type seems to be the best qualified to augment existing storage systems. It allows current data protection processes to be improved and automated instead of replaced.
The challenge with the third type is that specific integration to the storage system’s snapshot technology is required. It will take a few years for copy data vendors to provide integration to all the major storage system vendors. Copy data management or snapshot management will become available from a wide variety of vendors, but we think that only third party software developers will create solutions that will work across storage platform vendor products.
A Two Step Option
This “waiting for support” is why we felt it valuable to provide both webinars. The snapshot data protection strategy should be a first step in the process, it leverages the existing hardware without replacing it. It allows snapshots to be used for the most immediate data protection needs and rapid recovery of data from the most recent copy of data. Its weakness, finding and organizing data, can be offset by a backup application for now and eventually solved with a copy data solution.
We encourage you to watch both webinars. These designs can dramatically improve an organization’s ability to reduce both recovery point objectives and recovery time objectives today. They will also provide you guidance to how these strategies will evolve in the future.