Each year organizations are generating more data and keeping that data longer. This growth can be exponential, since traditionally many full backup copies are kept weekly, monthly, and yearly. For example: One file could have 25 copies if the most recent weekly backups are kept for 8 weeks, monthly backups for 10 months (covering the remainder of the year), and yearly backups for 7 years. Using this example, just 40 TB of new data would need 1 PB of storage. Mechanisms exist to reduce some of the common data, like compression, deduplication, and snapshots, but portions of the data will be unique or not easily reduced.
This compounding growth affects the cost of tape even more, since many of these data reduction mechanisms are not available, typically only native tape drive compression is available. In addition, many often overlooked issues need to be considered when using tape. Those issues include the cost of secure offsite storage, cost to periodically retrieve backups from storage to perform test restores as needed for compliance or business policy, and the risk of older tapes being unreadable due to age or incompatibilities with new tape drives.
Individually monitoring performance and free space on traditional RAID-based volumes/LUNs becomes a huge burden. The constant juggling and adding of new volumes/LUNs when capacity or hardware limits are reached consumes more and more time. Further, each storage change often requires changes to an application’s configuration. For example, adding capacity to a CommCell that is using traditional RAID-based storage requires many configuration steps on both the storage system and the CommCell. The storage system needs to be manually configured to provision the additional storage on existing or new volumes/LUNs. The CommCell Disk Library might need to be modified with additional mount paths or a new Disk Library created, which might require new Storage Policies to be defined. This could also require that Subclients be manually rebalanced across the Storage Policies. These time-consuming tasks are eliminated when PowerScale storage is used.
The inevitable and often overlooked hardware refresh every 3 to 5 years will more than likely take up many nights and weekends, data center resources, and budget with traditional RAID-based storage systems. This process often requires extensive planning, a complete re-evaluation of performance, capacity, and future needs out to 3 or more years to size the new environment properly. This new environment will then need to be stood up alongside the old to allow for data migration, and it will likely require one or more outage windows to cut over. Provisioning data center rack space, power, cooling, and network infrastructure for this type of hardware refresh can be expensive and time consuming.