Clearing Up Cloud Confusion

One of the complaints about cloud storage is that it is not very neatly defined. The reality is that I don't think it ever will be. There are too many use cases and let's face it, the term "cloud" is being used for everything, not only on the storage side but also on the compute side. As we discussed in our article "What is Cloud Storage" there are many sides to the term. Ideally, we should have reserved the term for providers of storage that you access through the public Internet, for which you pay by the capacity you are using on a monthly basis. I think that's what most people think of as "cloud." Unfortunately, we didn't declare what the term should mean from day one, and that lead to the confusion we have now.

The growth of the term stemmed from the access over the Internet use case. Cloud-storage providers needed a unique architecture that had to allow for a scalable NAS-like environment running on commodity hardware and storage. These needed to be managed as if they were one or as "one-like" as possible. The term "clustered NAS" is typically associated with that architecture, as is "global file systems" and "file virtualization." As with every other term, vendors each have their own take on how this "oneness" should take place, and there are pros and cons to each. Cloud storage providers also needed ability to easily or automatically move data internally and globally, as well as the ability to scale the environment quickly and easily. Again vendors interpreted the meaning of the term largely on what their architecture addressed.

As these architectures began to be developed, suppliers and users began to look at them for use beyond the public-access model. Many enterprises could benefit from an extremely scalable, easy to use, moderately performing storage tier. There was some value on the vendor's part to use the cloud term so this lead to the term "internal" or "private" cloud. Storage system suppliers responded to these needs and all looked for ways that their products could become part of the cloud infrastructure. A whole host of start-ups have responded to the need. Many legacy vendors already had a product that was similar to what cloud providers were looking for. For example, if you had been shipping a clustered NAS solution or a file virtualization solution, couldn't you make some claim to being part of a cloud rollout? Sure you could, and they did.

I like to think of cloud storage as more of a goal than a tangible thing. The idea behind it is to offer a highly-scalable storage environment that is easy to use and very cost effective. The components that you choose are largely dependent on your needs. If you are a provider it may mean using commodity hardware and storage, if you are an internal cloud user it may be compatibility with your current storage infrastructure, and if you are a user it will likely be more about what the software does for you than the technology.

The key for storage managers is to look at the cloud goals of improving scalability and driving out costs to determine which of these components will work well in your environment. You can choose to start with a new infrastructure altogether or add components to your current one to make it more cloud-like. If you don't like the term cloud don't use it. Call it that easy, scalable-storage thingy. There is also the use case for storage that is the back end to a cloud compute environment. We'll cover that in our next blog.