Data backup provides recovery options from data loss scenarios such as hardware failures, data corruption, human errors. Typical backup solutions enable RTO and RPO SLA objectives to be met. In addition, customers require data to be retained long term in order to meet compliance and regulatory requirements.
Both public cloud and tape offer the ability to store and retrieve data for such long term retention (LTR) use cases. Here are some factors to consider in the decision of whether to use cloud/tape/both.
| Criteria | Public cloud | Tape |
| Scalability | Cloud storage utilization can be scaled up or down based on demand. | The density and throughput of tapes has been on the rise with each new generation. |
| Redundancy | Public cloud vendors offer the ability of their storage service to be replicated to different regions across the world. The options are AWS Cross-Region replication, Azure ZRS/GRS, Google Multi-Regional storage. This geo-redundancy not only protects against localized downtime events but also enables data to be accessed by end users from the nearest data center across the world. | Typically, tape is maintained offsite in a single location. It is possible to store data on multiple tapes and store them in different locations. One has to then consider the tape infrastructure requirements, tape media, shipping logistics and physical location elements in order to accomplish such redundancy. |
| Flexibility | Public cloud offers compute, analytics, AI and many other services that can potentially be run against the data in the storage infrastructure. | Tape only offers storage of the data. |
| Security | With stringent key management, physical security, and certifications catered toward Government/federal agencies; security risks are being addressed by the cloud vendors. | With features such as WORM (write-once read many) and encryption, data on tape can be securely retained. |
| Search and analyze | Data sent to the cloud can be indexed and this provides the ability to retrieve data at a very granular level. | Data written to tape can also be indexed. Search can then be used to determine which tapes need to be loaded to retrieve the data. The time penalty incurred to load the right tapes typically is not a big concern in LTR scenarios. |
| API | Standard APIs provide ease of integration with cloud service offerings to build effective hybrid cloud environments. | SAS/SCSI/FC interfaces are needed to interface with Tape drives. |
| Cost consideration | Cloud vendors charge for the storage on a per GB basis per month and also a retrieval fee for data that is pulled out from the cloud. Depending on the type of storage redundancy used, the storage costs vary with tradeoff on access and transaction costs. | Tape storage should take the following factors into the costing model: Fixed costs: Tape drives, tape libraries, software Variable costs: Tape media, shipping costs, administrative expenses, Tape storage cost (per tape per month), Hardware and software maintenance |
Cohesity offers customers the flexibility of using public cloud or tape for long term data retention needs. The criteria above are meant to serve as directional guidance to determine the fit for a specific need. It is also possible that both cloud and tape may be used depending on the dataset that needs to be retained. As cloud vendors are in a race to lower storage service pricing, public cloud is getting more and more attractive for long term data retention.
