Loraine Lawson recently interviewed Judith Hurwitz, president of business consulting firm Hurwitz & Associates and co-author of the recently released "Cloud Computing for Dummies." In part one of the interview, Hurwitz explained why an integration strategy is important for moving to the cloud. In this second half, she discusses integration with the cloud, what you should and shouldn't move to the cloud and a prediction for 2010.
Lawson: Can you help clarify what the integration solutions are for cloud? I've heard about integration as a service, but when I talk with vendors, they don't call their solution integration as a service. Is anyone actually providing integration as a service, where they handle all the integration in the cloud?
Hurwitz: Well, I haven't heard of anyone who says that their product is integration as a service. I think what you would say is some of these platform as a service vendors are offering sort of middleware or capabilities that would allow you to integrate between different environments but I'm not sure you could just do it out of the box, if that was their implication.
Lawson: I'm thinking maybe the confusion comes in whether you're talking about infrastructure or software as a service integration.
Hurwitz: Well, when you're talking about software as a service, there is definitely an infrastructure component to it. You can have some software as a service components that are just a small service that allows you to do one thing and it's pretty straightforward.
But if you look at some of the bigger software as a service platforms, they're not simple.
Again, going back to Salesforce, there are thousands of third-party companies who have integrated their small application in with Salesforce so that you have Salesforce and you want to have something that calculates sales commissions. Well, that's not a part of the base Salesforce application; there's a third-party company that specializes in some sort of sales compensation models and determining based on different factors. What companies like that might do is use the interfaces to integrate their packaged application in with Salesforce. So, if I'm a Salesforce (client) company and say, “Gee, I really need this other capability,” I now buy it from a Salesforce partner.
You'll find companies that have a service, where before it would have been just a product. So I have the ability to integrate data across, you know, five different data sources. Now, before that I would have as a customer gone out and bought that, but now instead of buying what I need to use that for certain projects I will use that as a service.
Lawson: So, sometimes you have to go out and find someone, like if you're doing Salesforce, you have to find a solution that maybe will provide that integration.
Lawson: And then sometimes the vendors will package the integration?
Hurwitz: Yes, that's right.
Lawson: Is there any clear direction on which model is going to win?
Hurwitz: I think it really will depend on what your goal actually is. So like anything else, I don't expect there to be one solution for everybody; it's just not going to happen that way.
Lawson: Are there things that cloud is not appropriate for? Cloud, SaaS, platform as a service, none of it?
Hurwitz: Yes, there actually are. When we think about cloud computing, there's the idea that what you're actually doing is managing a workload. So, for example, where the cloud really makes sense is when you get economies of scale.
Let's say I have an old application that people in finance use four times a year to do some sort of quarterly reconciliation. It's an application that was built for this company 15 years ago, was customized exactly for their type of business and exactly for what they need. This is highly customized, it's an older application, it's used four times a year by four people. An application like that has no place in the cloud. Why would you want to do that? Because you're not going to get economies of scale. If you just wanted to get it off of your servers because it's only used four times a year, that's when outsourcing comes in.
On the other hand, you may have something like electronic mail, where that electronic mail is a fairly straightforward application. A vendor could write it once and scale it. You can optimize that application so it is so fast and so effective that you can support 100,000 customers or 2 billion customers, just by adding incremental capacity.
So applications like that are absolutely ideal for the cloud because one of the key factors about the cloud is it is an economic model. So just putting something in the cloud that is only going to serve those four people, is very complicated, it's not service oriented at all, you get absolutely no financial benefit. Does that make sense?
Lawson: Yes, absolutely. Is it important for the company to have a service-oriented architecture if they're going to engage in cloud computing?
Hurwitz: I believe a service-oriented approach is really critical, because if you look at it, any way you want to move away from monolithic approaches, you don't get the reuse, it's much harder to change. So the world is moving toward service orientation and if you think about cloud computing, in most situations, you're not just looking at sort of one approach and one application, one environment, you're really looking across things like what platform of services do I want. You want to make sure that the services are designed so as you change, as your environment changes, you're able to take advantage of new innovations, new platforms and, if it's designed in a service-oriented way, you have the ability to pick it up, it's a cohesive service with well-defined interfaces.
Lawson: Okay, I guess I'm going to squeeze in a prediction question. Any predictions for big integration-related topics in the coming year or trends?
Hurwitz: Well, I think what you're going to see is this notion of the hybrid cloud that is a combination of, “I'm going to use some public services and some private services,” so some services will be within my own data center, people will absolutely have a combination of clouds that actually live within their own environment and combining with clouds that are based on a public cloud infrastructure.
The other prediction is that brand and trust are going to be the only things that matter in whether a cloud provider is successful or not.
Lawson: Who do you think will have the brand and the trust to emerge as a leader?
Hurwitz: The companies right now who have the most brand equity are companies like Amazon and Google and Salesforce - companies that have been early pioneers and have really gained traction and gained trust in the marketplace. But I think what you're going to start to see happening, and you're already starting to see, is what I call the big elephant coming into the market. You know, the IBMs, the HPs, the Microsofts - some of the very big enterprise players looking for leadership.
And I think it's going to be a very interesting time.