I’ll be up front and clear. This post is an overview of Cloud Computing and why museums should care or at least be aware of it. First though, I find the progression of the server room quite interesting.
In the beginning one would have a singular huge computer that would crave the computational power of your cell phone. Access to these main frames was given to the end users through terminals or consoles. As servers became smaller and relatively cheap (< $10k), we began to fill our server rooms with many racks of computers. Each server typically had a dedicated purpose. Much is wasted with this model, since many of these servers would sit idly by +90% of the time while still consuming an immense amount of power. Not too long ago the Virtual Machine gave us the ability to run many virtual servers on a single machine. This was just a step away from where we are today. Why not run these virtual machines on someone else’s server?
Enter Cloud Computing. This buzz word should be in your repertoire for the next several years. It refers to running services and housing data on remote machines that you don’t manage. Dealing with hardware failures, upgrades, and redundancy should not be of your concern. The pay-as-you-use model is also much more affordable for smaller institutions.
There are two basic models to consider:
Software as a Service
This refers to using software API’s to run your software remotely. You don’t have any access whatsoever to the machines that are running your code. The advantage to this approach is that your app can scale up or down pretty much for free. Different requests to your service might run on completely different servers. Google App Engine is an example of this. There has even been signs that a future version of Microsoft Office will provided as a service.
Platform as a Service
This approach is basically providing you with the same functionality we currently have in our server rooms. Amazon Web Services is one of the most mature examples of this. They provide services to let you run your own virtual machine image, store your data, process credit card payments, and much more. You only pay based on the features and quantity of your usage. The IMA could technically replace its whole server room for only a fraction of the price. The advantage to platform as a service is choice. If someone else comes out with a cheaper or better platform, you aren’t tied down by an API.
The IMA is already utilizing some of this technology to enable itself to provide experiences that were not previously financially possible. Stay tuned…
Filed under: Technology