By Sramana Mitra and guest author Shaloo Shalini
SM: The 10% cost savings number is not what I am hearing from those who have really adapted to the cloud model. I have done a lot of these interviews. At this point, if you go through that series you will see people have been gaining a lot more than 10%.
DH: I am sure that there are people who will save may be 90% with cloud-based models. I’m also certain that this is a new frontier, and I am speaking in generalities here. There are many reasons today for moving to a cloud-based model, right? It could be as simple as saying, ‘I am lawfully in need of an infrastructure upgrade and I don’t have the capital expenditure budget to do it.’ Let’s say my application simply won’t run in the infrastructure that I have any more with changes in usage, scale and several other factors. There could be many other reasons. Well you know what, in case of such reasons, it might be a very good move if you go to the cloud. Now, if you want to measure it in terms of total cost of ownership that might not work out for you. But if you want to measure it in terms of, say, ‘return on investment,’ when you don’t have the capital of dollars to invest, it is going to be a great decision to think about cloud. I would say that the issue is the fundamental IT question – regardless of whether the IT model is cloud, client-server, or mainframe. All of these basic questions need to be answered. What we often find in such cases under the hype that has formed around the cloud, people are ignoring some of these fundamental tenants of good IT governance and just running toward a cloud strategy. It is all because of what vendors and all the hype around the cloud right now.
SM: Let me ask the question differently. This interview is for the Thought Leaders in Cloud Computing series on our blog. Could you give me some examples of areas where your work in cloud computing qualifies as thought leadership?
DH: Well, sometimes leadership means bucking the trend. Our leadership right now, our thought leadership – both internally and with our customers – is that if you need to solve a business problem and if cloud computing is a model that you think will work for you for whatever reason, then you should adopt cloud computing on a private cloud basis. I don’t define it the way you do. I define it as a virtualized infrastructure that supports utility-based consumption of technology inside your own four walls, and that is where you should solve for cloud computing paradigm in the context of your business problem first.
Understand in that context what are your integration risks, understand what are the associated provisioning and management risks, identify what some of the security holes might be in that model for your specific solution and use case. Another important aspect that you need to consider is, How can you pick and tie all of these data points back to whatever governance that is there in your data center, irrespective of whether that is internally imposed or imposed by some governing body. You need to be able to answer any compliance audit out there, and then and only then, look at whether it make sense for you to move some of your applications to off-premise. Once you have addressed these considerations, then maybe you could move to some cloud provider for burst capacity purposes. Another good reason to move to the cloud once the basic consideration are addressed could be because you want that economic model whereby you are going to pay by the drink as opposed to traditional cap-ex depreciation, op-ex for maintenance over the life cycle model, and stay on an upgrade path, which has been the traditional IT consumption model.
SM: When you talk to your customers, what are they saying about these issues?
DH: We hear a couple of things. We see customers with a lot of success and a lot of appetite for continued adoption of specific software as a service application such as Salesforce.com or such as Microsoft Business Productivity Online Access Suite (BPOS). We even see adoption of unified communications as a service. Customers are not opting to put in new voice infrastructure, and they are moving that to the cloud. It is mostly so that they can contain and control things. We also see a lot of customers with appetite for figuring out how they can create inside their own firewalls, inside their own infrastructure, a way to optimize and virtualize their IT assets such that they can provide a much more streamlined provisioning process. In a customer data center, when a new application comes on-board, when an upgrade happens, or when a new server needs to be provisioned, they need to do it in minutes or hours rather than days or weeks. Those are the two trends that we see happening every day.
SM: Okay. What other questions should I have asked you in terms of your cloud strategy? Are there any specific areas that I may have missed discussing?
DH: I think we agree that whether it’s an IBM or HP or Oracle or Cisco, that integrated stack is surely going to be an enabler of cloud computing. That is going to help a lot in the adoption process because of its benefits – it is pre-configured, pre-tested, and already highly integrated. I think that we agree on this detail.
SM: Well, it remains to be seen how applications can come on top of this because most of the vendors that are providing software service in the public cloud model today are not interested or willing to provide the same in the private cloud mode.
DH: Yes, that is one of the things. We have partnerships with a couple of what I would call public infrastructure providers, and we have engaged in probably half a dozen sales engagements. In all of those engagements, the customer has indicated a lot of interest in the provisioning, management, and the packaging that comes along with the public cloud infrastructure. But in the end, it always ends up in the same place. They typically say that they really like those offerings but would like the vendors to provide them such functionalities in their own infrastructure. As the major vendors get their act together and get their provisioning and management software more tightly integrated and then get that integrated stack tested with the major ISVs out there, and the major application providers out there, that is going to be a viable consumption model. When it is time for your next or may be first SAP deployment, the infrastructure blueprint is included; it is tested and now well understood. It is not only pre-tested, but after installation, you can have a complete life cycle of support for it. This is because the infrastructure is purpose-built for Microsoft Exchange or Lotus Domino or SAP.