Jim: How is IBM leading the cloud innovation?
Mac: I think we have done a good job of getting the market to understand that the world is a hybrid world. I remember the first analyst briefing I did with Gartner 4.5 years ago. Gartner’s definition of cloud at the time was what Amazon was doing. It was public cloud only. There was no concept of private or hybrid cloud. We really did a good job of getting the analysts and the customers to realize a lot of the value of the cloud, independent of whether it’s public or private or hybrid. We have done a good job of educating the market, if you will, and educating the analysts that this is a hybrid world. There is a place for private, there is a place for public, there is a place for hybrid, and a lot of our assets that we have for integration are vital to that hybrid world.
I also think we are doing well with big data and analytics. We have acquired many analytic companies, over the last 18 months in particular. Some of those analytic companies had very robust cloud-based analytics services.
When I talk to vendors and customers, it’s almost a standard response: “We know IBM is best at analytics, we want to understand what you have.” They do not say “Prove to me that I should pick IBM for analytics.” It’s the same way with integration: “We know you have CastIron, and some of the other integration assets, so we believe that IBM is in the best situation to do integration between public and private world.”
Some of our security services are very robust, but are not easily consumed. We want to take many of our assets that were built for a vertical tightly coupled architecture and transition them to be available in a loosely coupled, distributed, architectural model. Too often that puts me in a difficult position and is one of the biggest struggles we have.
Jim: Going back to the tiered architecture that many of our existing customers have now with the stack of products — how do we convince them or help them move to the cloud, and how do we sell to make best use of our time and theirs?
Mac: When I have the conversations with our enterprise customers I start with “what is the enabling technology for the cloud?” It is web services, which is why Amazon calls it Amazon Web-Services — because it’s built from a web-services model. I used to do service-oriented architecture (SOA) road shows all the time. What was SOA?
Basically we broke down tightly coupled components within an application, which were very customized and running in one place, and decomposed them into consumable services that could be integrated with other services from other applications. Basically cloud is taking some of those web services concepts and using them for IT.
Break down IT, compute, storage, network, and all the elements that sit up under each of those and make them consumable by way of web services. If you think about it, it’s an extension of the SOA revolution — it started with the application. Now that you have the applications so they can be agile, and plug and play, and then integrate things together, all of a sudden the infrastructure is in your way. Now you want to make that agile, it’s just the next generation of web services enabled. We are basically enabling a different tier in the customer’s environment — before it was the application, now it is network. To me, you choose what the best fit is for your organization. I try to help customers realize that the cloud is not something completely foreign and different, and they do not have to start from scratch.
Jim: Now as far as multitenancy versus single tenancy, have you seen much pushback from customers, and how do we get over this?
Mac: That’s a great question. This is a hotly debated topic in the cloud world. I can remember in the early days when I was working in the cloud, statements like: “If you are not multitenant you cannot be a cloud.” But if you think about it, some of the services running in Amazon’s cloud have been successful for some time. They were services built on the enterprise, that were single tenant and they used virtualization, and things like cloud formation as a deployment engine on the Amazon side to deploy a topology.
So I believe with the work we have done with the PureSystems, with the IBM Workload Deployer, we have provided some really good solutions, in terms of helping customers enable those kinds of born-on-the-enterprise workloads, which tend to be single tenant.
For example, I am a customer, and I am building a custom application, and I am going to consume that custom application with my known clients; they are all my users, so I do not really need multitenancy for that. I just need to get the compute, storage, and network resources, so that I can deliver the application.
But perhaps I am a software as a service (SaaS) vendor, and I am going to build an application that can be consumed by anybody, no matter where they are, or how many clients there are. If Joe the plumber from Indiana comes in and wants to buy my service, I’ll sell it to him. It’s a completely different model. Then all of a sudden I might start out with a single-tenant model, where I am spinning up VMs and attaching these users to that VM, and those users to this other VM. It becomes a very cost-ineffective model for me, especially if I am renting capacity from a cloud service vendor.
Often, we see ISVs and solutions that are born on-premises start out with that single-tenant approach, and then realize it’s not a cost-effective model if they will be doing a SaaS. If they are a custom application deployment, that’s fine, but to place where I will have multiple consumers in multiple locations, I do not know how to prepare for the kind of scale I will need. Therefore, I do not know what the cost structure will be; I need to be able to enable multitenancy.
Jim: Thank you for your time.
Mac: Thanks, I appreciate it.