The IT industry is using the term cloud pretty liberally these days. I'm sure you have seen the latest Microsoft "To The Cloud" commercials. I am not sure they are really helping people get this whole cloud thing. To that add the various cloud storage services as well as the cloud compute services and there is a lot of confusion when you let a CIO know that they can have a cloud right in their own data center.I am going to try to define this Cloud Data Center concept and while I'm sure there are others I think this makes sense. First, to level set, let's remind people of cloud compute and cloud storage. Cloud compute is the term typically applied to a situation where you are going to borrow someone else's processing capabilities most often over an internet connection. It may be to host your own application, to start a group of servers to simulate a condition or it can be your use of an actual cloud application like a financials package or a sales automation program. Basically it is server resources or applications as you need them.
Cloud storage is when you are going to borrow someone else's storage for a period of time again typically through an internet connection. This can be to store backup or archive data. Increasingly though, as we will discuss in today's webinar "Adding NAS To Your Virtualized SAN By Leveraging Cloud Storage" it is being increasingly used for primary storage. This of course makes the use case more than just borrowing storage but it is certainly a case where the bulk of your data may not reside in your four walls. In each case it is the concept of buying or using storage as you need it.
A cloud data center takes these concepts and moves them in your data center. The goal is to bring the flexibility of the "as you need it" resource allocation to your own data center. The vision is to move any resource; compute, network, storage, etc... to where it is most needed at the moment it is needed, with a minimal amount of effort, best case being automatically. The cloud term is applied to portray the abstract nature of this data center, that components of it can be moved around easily. I agree the cloud term is a bit of a stretch but I think we are stuck with it.
The major, by today's standard, component of this cloud data center you probably already have; server virtualization. The ability to deploy a virtual server on almost any piece of hardware as you need it. That is abstract and for many vendors qualifies you as a cloud data center, but compared to our definition of any resource at any time, it is really just the beginning. Where the cloud data center today falls short and where much of the work is being done in completing this vision because in many data centers moving a virtual server from host A to host B is not as simple as a point and click nor is the ability to sub-allocate the other resources in the environment.
The ability to move virtual machines to other servers brings networking and storage challenges with it. As we discussed in our recent article "Network Limitations to Cloud Computing and Network Convergence", in many cases when a virtual machine is moved the network needs to be reconfigured, manually, to match the network configuration it had on the prior host. Network ports need to be able to move virtually with their virtual machine so no reconfiguration needs to take place.
Storage is another area that needs to improve to support and help complete this vision of a cloud data center. We'll cover some of those issues in an upcoming entry.
Track us on Twitter: http://twitter.com/storageswiss
Subscribe to our RSS feed.
George Crump is lead analyst of Storage Switzerland, an IT analyst firm focused on the storage and virtualization segments. Find Storage Switzerland's disclosure statement here.