What exactly is cloud storage infrastructure. That question came up recently when we began a project to identify some of the top cloud storage infrastructure providers. The goal is to provide organizations looking to build their own private cloud or service providers looking to provide storage as a cloud service a starting point when looking for the storage infrastructure that will support that effort.
In brief, cloud storage infrastructures are the hardware and software (or both) that are either assembled or delivered turnkey to physically store data, either for an organization building a private cloud storage system or a provider that offers public cloud storage services. These are not onramp type of systems that help get data to cloud storage--they are the services and systems that help provide storage in the cloud. These systems should have some key capabilities that we will look at over the next few entries.
First, there should be a component that can be either fully or partially installed in your data center. In most cases, when we think of private cloud storage we think exclusively of on-premises equipment. But as we discuss in our recent video "What is Hybrid Cloud Storage?," some public cloud providers are now able to extend the public cloud into the data center, which in my mind now qualifies them to be an infrastructure provider. In either case, on-premises deployment is designed to reduce latency and maintain some sense of control.
Second, cloud infrastructures have to scale. When we talk about scaling the cloud infrastructure, we typically focus on how much capacity the storage environment can handle with a single management point. These systems all have to be able to scale to multiple petabytes and all the data should be accessible from a mount point. I purposely don't use the term "file system" since many of these systems are object based. Essentially, they should have a single container that can scale to multiple petabytes of storage. The goal for a cloud storage infrastructure product to eliminate storage system sprawl as much as possible.
Capacity isn't the only issue when it comes to scaling, and some of those issues might force an organization to implement multiple systems. One is a limit on the number of files or objects that can be stored. This could be a problem encountered early on by service providers that hope to have thousands, if not millions, of users all storing various types of data.
There is also a performance aspect when it comes to the size of a cloud storage infrastructure. If the cloud storage infrastructure is being used to support a repository function like archive, file sharing, or backup, then modest but consistent performance is needed as capacity is added. These are the systems that our project is focusing on.
There is also a cloud storage infrastructure that is going to be used to support cloud computing applications and in that case, performance demands will outweigh capacity demands. That issue will probably be a separate project for us in the future.
Having an on-premise capability and the ability to scale are just the initial capabilities to look for in a cloud storage architecture. Another is the ability to curtail costs and is something that we will look at in the next entry in this series.
Follow Storage Switzerland on Twitter
George Crump is lead analyst of Storage Switzerland, an IT analyst firm focused on the storage and virtualization segments. Storage Switzerland's disclosure statement.