Loraine Lawson spoke with Judith Hurwitz, president of business consulting firm Hurwitz & Associates, about her research on cloud computing. Hurwitz is the co-author of the recently released "Cloud Computing for Dummies."
Lawson: Your work has been focused on cloud computing recently, is that correct?
Hurwitz: Sure. We just came out three weeks ago with a book called "Cloud Computing for Dummies," which is very exciting. We researched the topic for almost 2 years, so, yes, we have spent a little bit of time with it.
Lawson: Just to clarify, are you talking about hardware infrastructure here or are you also talking about SaaS?
Hurwitz: We look at cloud computing as three types. There’s infrastructure as a service, and infrastructure as a service means that you're basically buying capacity. It could be storage, it could be CPU cycles, it could be any type of incremental service and that’s typically what I call sort of a “bare metals” platform. You're buying capacity when you need it. You might even put your own operating system, your own software on it. You may use it for things like development, you want to develop a piece of software, you want to put all of your own tools on that, this way you avoid having to go out and buy a lot of physical resources to support that effort. You may use it for testing an application. You want to see what happens when 10,000 customers all try to use a service.
“I think the other issue that has to be said here is that we are relatively early in cloud computing. It’s not a mature market. So there’s a lot of tire kicking, there are a lot of experiments going on.”
- Judith Hurwitz
- President, Hurwitz & Associates
An example of infrastructure as a service might be Amazon EC2 cloud. If you are needing to get some extra computing capability, you may sign up with them and you can then get provisioned to you an extra terabyte of storage that you need for a project which probably will go away in a month, but it allows you to do that without investing in physical infrastructure. So that’s one type.
Going a level up, there’s something called platform as a service where, in addition to getting access to certain capacity on demand, it includes a lot of other services. So it will have an operating system, it will have some sort of middleware, it may have software development tools, testing tools. It may have a whole package of infrastructure that is all integrated into this one platform. So that’s platform as a service.
Platform as a service might be something like Salesforce.com’s Force.com, which is a whole development environment including languages and all sorts of tools and it allows their partners to go and build a new application that takes advantage of all their basic services that they themselves use to build their software. That would be a platform as a service.
And then the third, and really where this all started, is software as a service. Salesforce.com is really the example that we all think of when we think of software as a service because they were really one of the pioneers of this area. And that sits on some sort of platform. It offers capacity on demand, so if you are using Salesforce as your sales management platform and you add 500 new sales people tomorrow morning, you don’t have to say, “Okay, what other hardware and software do we need to buy to support them?” As a service, you just expand without having to worry about buying extra hardware resources.
Lawson: It sounds like companies are using infrastructure in the cloud for temporary needs. Is that what you found?
Hurwitz: Yes, it’s actually quite common. For example, it may be different companies are doing a joint development project together or it may be a small team. They may be even experimenting: “Let’s see if we can go build something like X,” and you know, you're not ready to commit to it on a corporate basis, but it would be useful to see is this the right thing, let’s do a prototype.
So rather than saying, “Okay, to do this prototype I want you to give me all of these resources,” what you do is you go to Amazon, you open an account, you give them a credit card, you say, “Okay, let’s start building this,” and they will charge you 10 cents for every CPU hour that you use. Then when you're done with that application, you take a look at it, you show it to management and then people decide, “Yeah, this looks like exactly what we need, now I’m going to put it into production. I may keep it, I may decide to keep it on a cloud platform and make that economic decision,” or at that point I may say, “Okay, this is great, now let’s implement this internally.”
People make different decisions, but it’s a very effective way to do something quickly - I just want to get started, I don’t want to worry about any infrastructure issues.
To ShareThis, click on a service below: