Cloud providers like Amazon, Google and Azure all provide both compute and storage services. Compute is used to host applications and storage is used to store data. The cloud is tailor made for providing compute because applications that are multi-threaded can easily consume or give back resources as needed. Providers though struggle with storage, because data has gravity, and is hard to “give back.” In addition, most cloud providers’ storage offerings are woefully lacking in the data services and reliability applications demand.
The Challenge with Cloud Storage as Primary Storage
Primary storage systems have three responsibilities. First, deliver the I/O performance that applications demand. Second, protect that data from a media failure, data corruption or accidental deletion. Third, provide a robust feature set that allows IT administrators and/or application owners to effectively manage the environment.
Each of the big three providers offers primary storage for applications they host and in most cases the performance of that storage is not the top concern. Where the primary storage solutions that cloud providers offer fall short is in the area of reliability and data management. From a reliability perspective, providers typically offer three nines of uptime, while most enterprise systems over five nines or more. For data services most providers offer very limited capabilities, even snapshots are hard to fully implement.
The other challenge with using a cloud provider’s internal block storage offering is that the organization is, at that point, almost totally locked into that cloud provider. As application portability continues to become easier to implement organizations are going to want the ability to move applications between on-premises and multiple cloud providers. Settling on a provider’s block storage offering doesn’t make that movement impossible but it does make it more challenging.
Introducing Nimble’s Cloud Volumes
Nimble’s Cloud Volumes takes the cloud primary storage challenge head-on. Essentially a Nimble storage system is implemented inside the same data center as the cloud provider, or in close proximity to it. Organizations with cloud hosted applications can then connect their cloud account directly to the Nimble Storage system. The result is cloud-hosted application enjoys all the performance, reliability and data services of an enterprise class storage system.
Cloud Volumes, though, is more than just a physical placement of an existing product. Nimble provides a new front-end to their storage system that for this use case is greatly simplified and better matches the self-provisioning portal that cloud based customers are used to.
Cloud Volumes is also priced as a service. That means customers don’t have to pay upfront for their storage system as they would with a traditional Nimble array. Instead, it is a more cloud-friendly consumption-based payment model.
In the future, existing Nimble customers will be able to replicate their data to a Cloud Volumes device which will provide an ideal disaster recovery solution as well as a means to fully exploit hybrid-cloud bursting.
Any time we see a model that exploits cloud compute without sacrificing data security and management, we are interested. Again, data has gravity, CPUs are expendable. Cloud Volumes provides organizations with a path to the cloud without sacrificing data reliability and manageability.