By Sramana Mitra and guest author Shaloo Shalini
SM: Based on my research, I see a distinct trend [toward] interesting private cloud deployment.
LO: Very interesting! What would you say is the driver; I am curious to know.
SM: I think it’s cost that is the driver, which is why I was probing that, to understand what your perspective on private cloud adoption was. To some extent, it has happened in some cases that cost is the driver. Earlier, you mentioned cases that required the kind of data security and clients did not want to co-mingle their data with other data in a multi-tenant multi-cloud environment. Those are some of the drivers for private cloud adoption as well. What surprised me was – and I have heard it from SMEs as well – that there are the large vendors of infrastructure who are pushing private cloud’s stacks to their large customer base. It is they who are trying to build analytics clouds internally.
There are certain applications where they want to have a more private infrastructure, say for I/O intensive applications where they wanted to do private clouds. There are a variety of drivers for private cloud adoption, but I think that cost is also a major driver. I think in some cases, maybe they don’t want that highest common denominator of cloud-based services. Maybe they want something smaller or less expensive, and the cost issue is significant enough for them. By going to a private cloud architecture, if they can save 30% of the cost, they would be willing to do it. Having said that, most cloud vendors are not set up to deliver a private cloud solution today.
LO: That is true. You end up almost running a traditional perpetual licensing or packaged software solution in a dynamic elastic infrastructure layer, right?
SM: No. Well, yes. But I think the pricing models they are looking for are still the subscription-based pricing models. That is what you are doing with the Department of Defense. You are still following a usage-based pricing model there, so you still get the benefit of the usage-based operational expense pricing model.
LO: It is interesting.
SM: There are about 600 SaaS application vendors right now, and most of them are not set up to deliver anything in the private cloud. They are all in a public cloud mode, but it is just something I have seen coming from the customer side. This is one of the reasons why we are publishing the TLCC series and getting a cloud adopter’s point of view. This series is focused more on taking into account the cloud customer’s perspective, not the vendor’s perspective.
LO: Right. The DoD deployment in a private cloud environment was an interesting exercise for us. Although their primary driver was not cost, it was more of security and sort of what they required, it was interesting as a cloud provider to package that solution that way in a remote environment where we didn’t manage the infrastructure because that is typically what we do; we are managing the full stack.
SM: Switching directions a bit, from where you sit today, in terms of cloud-based solution adoption, where would you point entrepreneurs to look for problems to solve? Where are the open problems, where do you see blue-sky opportunities that you would send entrepreneurs after?
LO: You are speaking of entrepreneurs who are looking to provide cloud solutions and services?
LO: I think one of the areas where the cloud can indeed provide a lot more value is on operational management services. I kind of talked about risk management earlier. There are some significant gaps, I think, in the industry there. As an industry, we spend so much time focused on feature functionality and trying to catch up on the application maturity side and sort of providing that aspect of value. But we have left some of the IT stuff off the table. Well, certainly right now we have been trying to do that. I’m speaking more as a CIO in the context of cloud solutions and the ability to consume operational services, whether it is monitoring or security. Compliance, standards, and those types of things could do a lot to burnish the image of cloud computing in the corporate enterprise. I am certainly connected with some people who are doing entrepreneurial work across the gamut – how to help manage risk plus a variety of other things.
It is certainly one category that would be important in operation transparency. Say intermediaries, right? Integration and intermediation are two of the areas that people get concerned about [with] computing. The concerns are related to user lock in and whether I have to worry about, say, ending up in APEX and whether I could get stuck. Say I wrote all my code there; now I can’t go anywhere with it. That is another area where I think intermediaries would be required, whether it be integration intermediaries [or something else], entrepreneurs could help eliminate or deal with some of that lock in. That could be an interesting opportunity for entrepreneurs