Last July, Alistair Croll wrote an interesting post over at http://datagravity.org on the concept of “data gravity” – a term first coined by Dave McCrory. The concept goes like this: The speed with which information can get from where it is stored to where it is acted upon (e.g., from the hard disk in your PC to the processor) is the real limiting factor of computing speed. This is why microprocessors typically have built-in cache memory – to minimize the number of times you have to go back to the storage repository to access the data that a running application needs. What’s interesting is that this has practical implications for cloud computing.
Microsoft researcher Jim Gray, who spent a lot of his career looking at the economics of data, concluded that, compared to the cost of moving bytes around, everything else is effectively free! Getting information from your computer to the cloud (and vice-versa) is time-consuming and potentially expensive. The more data you’re trying to move, the bigger the problem is. Just as a large physical mass exhibits inertia (i.e., it takes a lot of energy to get it moving) and gravity (i.e., it attracts other physical objects, and the larger the mass the stronger the attraction), a large chunk of data also exhibits a kind of inertia, and tends to attract other related data and the applications required to manipulate it. Hence the concept of data “gravity.”
The implication for cloud computing is that, if you’re going to move your data to the cloud, then it only makes sense to have the applications that will access and manipulate that data in the same cloud, and vice versa. Otherwise, you’re going to constantly be moving data in and out of the cloud repository. And, of course, that’s what DaaS (Desktop as a Service) is all about: You can move your data to the cloud, and you can get a cloud-based desktop session in which your applications run. Your data and your applications in the cloud – together, as the concept of data gravity suggests that they should be.
And, of course, you get additional benefits, such as the ability to work from any location where you have an Internet connection from any device that will run the Citrix Receiver client (assuming your DaaS provider leverages Citrix technology), you have an instant DR/BC plan, you get fast and easy scalability, and you can immediately support a BYOD (“Bring Your Own Device”) initiative.
Don’t fight gravity…
One Thought to “Data Gravity and Cloud Computing”
[…] the same time, we wrote about the concept of “Data Gravity” – that, just as objects with physical mass exhibit inertia and attract one another in accordance […]