In our on demand webinar “How to Design Self-Protecting Storage and Gain Backup Independence” Storage Switzerland and ClearSky Data articulate how primary storage can be improved so that it can take on more of the data protection responsibility. The goal is to enable organizations to become less dependent on traditional backup and recovery solutions. The first step in that process is to understand the limitations of typical storage systems when trying to use them to protect themselves.
Data Protection vs. Backup
Most modern primary storage systems today do an adequate job of providing data protection. Through the use of RAID or erasure coding, they protect against one or even multiple drive failures. Scale-out systems, again leveraging erasure coding, even protect against node failure. These systems, today, also provide snapshot capabilities so that for a limited period of time they can “roll-back” volumes to a specific point in time. Finally, they provide replication to move data off-site which protects against a site failure.
The data protection provided by most primary storage architectures can’t qualify as backup because most of the protection features are too limited in scope and have too short of a retention. Media level protection like RAID or erasure coding, while critical, are too limited in scope as they only protect against one thing, a failed drive, not data corruption or rogue users.
Snapshots save their data to the same storage system as the original volume which limits their scope too. A problem with the entire storage system or a data center level disaster will likely destroy those snapshots. Snapshot copies also grow in size as more changes happen to the primary volume. Again, because they store data to the same storage system as the primary volume they are also too expensive to use for long term retention requirements. Finally, many storage systems will suffer a performance impact as more snapshots are taken and those snapshots age. Updating all the various copies can dramatically impact performance.
Replication is similar to snapshots; in fact many systems use the same technology to perform both functions. The key difference is that replication moves the data out of the primary storage system, to another storage system. Some systems allow separate snapshot schedules on the secondary system which alleviates some of the challenges with snapshot based protection. The problem is that in the large majority of cases the secondary system must be the exact same configuration as the primary system which doubles the cost of the infrastructure. It also makes the secondary storage equally unattractive if the organization wants to use it for point-in-time recoveries. Additionally, the organization typically will only replicate once and the replication target, to meet disaster recovery requirements, is placed far away from the primary. This distance makes the secondary system less than ideal to service day to day recoveries.
Primary storage systems provide excellent protection as long as the problem is caught quickly and they also provide very rapid recovery. The problem is they don’t offer adequate protection from a problem that takes more than a few hours to identify. For these situations, organizations need to either add backup or look for a truly self-protecting solution that delivers longer term point-in-time protection as well as data retention.
In our on demand webinar “How to Design Self-Protecting Storage and Gain Backup Independence” Storage Switzerland and ClearSky Data detail the challenges that traditional storage systems face and how a cloud-based primary storage system provides full protection while enabling the organization to keep its computing assets on-premises.