Sramana Mitra: What do you consider being open problems right now? This is one of the areas in which our audience is always very interested. If an entrepreneur is looking for open problems to solve right now, where do you see opportunities?
Mark Settle: That is a good question. Most of the large enterprises are dipping their toe in the water. They are exploring cases like DBi in the cloud, where a virtual desktop is offered. Maybe they have contractors or training facilities in different parts of the globe. I believe that storage in the cloud is a real area of opportunity, and I know investments have happened there in the past. But storage continues to be a large cost factor for most organizations. I think there are lots of ways to make the cost of storage more transparent by getting a monthly bill for it, and then dividing that bill and charging departments back. This is opposed to the classical way IT usually makes business, which is to creep more and more storage over time and then charge it back on a block basis. You can make this much more transactional with the use of a cloud provider. I think this offers a lot of opportunities.
The management tools that I see in the private cloud sector – the people who have been successful had real niche capabilities – over time those will be swept up into broader management platforms. It is probably the convergence of where the private clouds come together with the third-party cloud. What are the right trigger points, economically and operationally? When I take a job or a task and I decide I want to do it inside the firewall or outside the firewall … one are that merits a lot of attention is workload management. We talk about workloads – managing them in a private cloud versus managing them in a public cloud – but I believe we are using the term incorrectly. What I mean is that people think of a hardware perspective when thinking of a public cloud; they think: “How many VMs do I need or how much storage do I need?” And they basically want to rent or lease somebody else’s VMs and storage to start doing what they wanted to do. I wouldn’t be surprised if we are heading towards longer term. If you have a specific workload, which you are trying to perform internally, and we would be able to characterize them in a specific fashion, you could basically reverse auction off those workloads to third-party providers. Then you could prepare your internal costs as well. If I was going to market a basket of 500 stocks and I wanted to do that every five minutes, you could look at the utilization pattern within my own private cloud and decide for yourself if it is more cost effective to for me find more computing capacity in the next five minutes, or should I go to Amazon and rent [it] for the 30 seconds it is going to take me to do it. You could actually have a marketplace where you can advertise the workloads you are trying to have performed. They can even have security designations. You could have a workload which is very security sensitive, which bumps out something internal to the firewall, which you do on a routine basis. So you take the routine workload and let that go out with a third-party provider.
SM: How would that play out? Would the vendors have to look at these offers and then bid for them, or do you see an automated bidding machine?
MS: Somebody could intervene and create a marketplace for that or take the features of some of the public facilities. If you look at the economics of this, the public providers are making big capital bets as they continue to expand capacity. It is like in an airplane: you don’t want any seat to be empty, even if you sell it at just a fraction of what it would cost to move a passenger 900 miles. That seat is going from New York to Los Angeles regardless of anybody is in it or not. So, you have all that capital just sitting there. You don’t turn machines off – nobody does that in a data center. Whether it would be a third party exchange – larger enterprises would advertise their needs – it would turn the table on the public’s eyes. They would actually be bidding for the workloads. What we lack today is a nomenclature for categorizing these workloads. I am talking about a common format that allows for competition.
SM: It is interesting that you mention this marketplace model. It is one of the biggest trends I have seen in the past three to four years. I have done countless case studies on the various marketplaces, and I would say the most active one is the services marketplace trend. People putting projects up on sites like Elance, oDesk, Freelancer.com, or 99designs. People from all over the world are bidding for them and winning services. I think this is going to become a trillion dollar industry.
MS: I agree. The potential for it is definitely there.
SM: Would you say there are similar opportunities in the infrastructure space, computing power or storage?
MS: Yes. Let’s start with the premise that nothing is sacred anymore. In the old days I would have said, “Storage in the cloud will never take off because there will be all those security concerns.” Admittedly, there are certain industries which would be reluctant or slow to adopt that concept. For example, healthcare providers, which store a lot of personal information, financial service firms, etc. But then there are other industries for which it would be much easier to embrace those technologies.
SM: This sharing of spare capacities is happening in other industries as well. One that comes into mind right away is the hotel industry, which is getting hammered by services like Airbnb, where people who have spare rooms or even a couch are offering those up for rent to travelers. People are using Airbnb instead of booking a hotel room, for a fraction of the price. This is causing hotels to sit up and take notice. The same thing happens in the car industry. One example is Zipcar. Especially in urban areas, people are deciding not to buy a car and when they need one, they rent a Zipcar. This is a major trend which the entire car industry will take notice of and will have to see how it impacts their world.