Milestone Group Quarterly: July 2005
Articles
Face to Face: Bill Coleman, CEO of Cassatt
Milestone: Tell us about the genesis of Cassatt? What’s the big idea?
Coleman: Well the big idea about Cassatt was looking at the emergence of this decade. Seeing that what’s happening with service oriented architectures and with greater utility computing is a commoditization of computing. My “aha” was the complement of the sort of same epiphany that I had to form BEA, that the network would need an operating system. In this one, I believe that the network is going to be the basic computer infrastructure that needs an operating system. The telephone switching system, be it an analogy or not -- what it did for the telephone system eighty years ago -- we have to do to computing. We have to take a holistic focus.
Milestone: A lot of people see utility computing and virtualization evolving slowly. Starting with today, what are the phases of adoption as utility computing becomes reality?
Coleman: I don’t think utility computing will be a reality until the end of the decade or the end of the next decade. The real phases are not the utility computing model as people think of it, where you are buying the service from someone externally. It’s the utility computing model internally. Where a company turns their own capital resources into a utility and are able to charge out for those. So that is the first phase, the adoption of utility computing internally. Then as the model evolves, it will be the commoditized players who I believe will win in the utility computing model. That means some evolution of what are today the Telco service players, the ISVs, and the portals. In my own belief, 15-20 years from now, there will be less than 20 of those in the world that will provide the 80/20 rule, 80% of all computing services. It will be totally commoditized. In order to do it, they will have to sell it as a utility, not only just by time of day, but also by capacity and demand. They will also have to sell it commoditized. In other words, if you can’t eliminate the OpEx costs significantly, which is the business driver, it will never work. It will be too expensive to provide. Once you can, you not only kill the model of high end computing today but also today’s outsourcing models, because today’s outsourcing model is value priced based on what a company can afford to do with high OpEx and CapEx costs today. Utility computing will be just the opposite.
Milestone: Novell, Sun, IBM and others are also doing similar things with virtualization. Friends or foes?
Coleman: I would look at anyone who is out defining a new market as a collaborator, if it is truly a new platform, which I believe this is. This is the new run time for IT automation because they are all validating the market. Basically, one of the characteristics of the new platforms is that it changes the economics of the business. So in this case it changes both the economics of hardware and it also changes the economics for the customer. So, the customer value of buying the new platform is to commoditize the economics of their current suppliers. Therefore, if you look at any platform in the last 25 years that has formed in computer hardware or software, it took a new entrant, a start up company, to define it. It doesn’t mean they will own it past a billion dollars, but at least to define it and then the others can become good enough.
Milestone: How does open source play, if at all, into the virtualization market?
Coleman: Open source is going to play in all software markets going forward. Open source, to me, has turned into the first viable economic model of reusable code. In our product we have 43 different components of open source. It accelerates innovation. It makes your spending of capital more efficient. It is much more efficient than going to India to do it. It’s already there and it’s free. It lets you focus on where your competitive advantage is, but it means it’s going to accelerate the commoditization of new and old product lines.
Milestone: So what’s the ROI on all this in a large enterprise if really implemented full scale? What”s the order of magnitude of cost savings, and what are Cassatt customers seeing?
Coleman: So what is the ROI of turning your system into a commodity, into a utility? With software like ours that allows you to run many applications on one grid, let’s say? In doing it, you can lower the cost, you can increase utilization. Our assumption is ultimately 75% of your OpEx costs and 70-80% of your CapEx costs. Pfizer is our first big deployment. They estimated on the first 70 applications they put on this environment that their three year cost savings was $25 million. That is 10% of the applications they will put on in the next few years.
Milestone: We’ve seen other virtualization type technology, for instance with the grid players, and it’s been slow to take off. How is Cassatt different?
Coleman: Cassatt is different from other players approaching this market in the same way that BEA was to begin with. We are not approaching this market as a technology; we are approaching it as a customer’s problem. Therefore, this utilization is only one of three tools that we happened to put together that eliminates most of the operation cost, enables commoditization of computing, and allows you to guarantee quality of service and do it dynamically. It’s a policy management framework, an office management framework and we don’t virtualize the physical world, we virtualize the logical world. The logical world is applications and files because that is all that flows around the network.
Milestone: You had an interesting approach to building the Cassatt team in the beginning. Tell us how you did it.
Coleman: My philosophy on building a new platform company, especially mission critical software, if the team has not done it before, is that the technical team must have been together for at least five years, this being the third version of the technology you ask them to build. If the management team has not done it before, you are not going to be able to succeed the first time. The management team you have to build top down. In order to grow fast, you cannot be constantly changing or upgrading the management team, because you will never have a culture or a set of values. Great people hire great people and that is how you can build an organization fast that is built to last.
Milestone: Not everyone can go buy a company to build a team, nor easily find entire teams with 5-years+ working together? Any tips for the rest of us?
Coleman: The first tip is start with your scope. Figure out what the core management principles are. There will be three or four areas of expertise that you really need, like a CEO, maybe a CTO, someone for products, someone for channels, someone for “G&A.” My view is that you are looking for the following sets of experiences: You are looking for someone that you can work with that you believe has the same sort of values that you do. Someone who is great at what they do, that is the second. The third one is that you are looking for someone who already knows how to manage in a large company. So he should have been an executive in a large technology growing company. You are looking for someone who understands the limitations and strengths of a start up, so they have to have been in a start up. Finally, it is very desirable if the core management have all been through a failure. Whether that is a total failure of a company or a big down turn because that is where you really learn what matters. You don’t learn what matters when you jump in and everything is up and to the right.
Milestone: Switching gears a bit…. Bill, you Co-founded BEA, and took it to a leadership position in the application server space. -- Incredible growth, in fact the fastest a software company to hit $1B in revenues from inception. Does lightning strike twice for you?
Coleman: I didn’t come back because I didn’t think there was another platform opportunity. A platform opportunity, as Geoffrey Moore told me when we started BEA, is to find something that people stand on. It’s not a platform if people do not stand on it. So what are you building? This is, I believe, a run time for IT operations. It is not systems management, it’s something totally new. Its goal is to eliminate the cost of operations, while enabling real-time business agility and guaranteed quality of service, and in doing that, there is a dramatic cost savings. Yes, it is a platform because policies for operations are defined and built into the platform. Your end user customers are configuring it. If that happens, it means Darwin is going to select one company to win and define this space and that company has the ability to get to $1 billion in revenue. Can it strike twice? Only time will tell, execution is what it is all about and a little bit of luck, which is what we had at BEA. But that is where I am squarely focused.
Milestone: Speaking of BEA, what’s your current take on the application server market? Mature, or still room for growth and innovation? Open source players walk away with the prize?
Coleman: I think the definition of platforms, and you can look at all the companies that define different levels of platforms, whether it is Microsoft or Netscape or Veritas or BEA, a company gets to build it and Darwin selects, as I said, to be able to give you the runway to get to $1 billion in revenue. That sort of seems to be the point in software. At that point you have to move beyond your current space. The signal to that change is when technology innovation is not the differentiator. You will first tell that when the mature spaces become good enough. When you start hearing that phrase “these guys are good enough,” you need to be expanding your footprint into another space. So, one, we have that market that is now being surrounded by people who are “good enough.” So it is not just the deployment platform for SOA anymore. The second thing accelerating this, as I said earlier, is open source. Open source now squeezes companies that are in the market from one side while the large gorillas are squeezing the other side. There is really a market for an independent, but at this point I think the decision is you have to widen your footprint and figure out some strategy to get to $2 billion beyond the web application server market and its directly related add-ons.
Milestone: Some people see BEA as not fulfilling its potential in spite of reasonable financials. You have been quoted as wanting to see some more aggressive actions. Like what?
Coleman: Still being a consultant for BEA, I don’t think it would be appropriate to give my ideas. But if you look at what is happening in the marketplace, think back historically. Through the four generations we have gone through in the last 35 years, successively building platforms from the chip processor level to the PC workstation level and the network and the software associated with it. It is interesting that the end of each decade was where the innovation happened. We had a down turn at the beginning of each decade, went through disillusionment. Then you had about 5-7 years where two things happened. The whole product was defined, which is happening now for SOA, and second, most of the firms that were in it in the beginning of the decade were consolidated out of the business. This is the point where you have to consolidate because at the end of the decade, if indeed it happens again; there will be a few players that will be left that will be considered buyable to build the SOA world on. I am on the Symantec board; you can see what I believe from that, it is a very viable strategy. I think that what John Thompson has done is put together the two biggest pieces of infrastructure that synergistically give the best leverage, and I believe this is as fundamental for future computing as the merger that created Cisco back in the 1990s.
Milestone: If you had to run a company in another space, what plays do you like right now?
Coleman: Well the first thing is I think there is a whole missing SOA platform where I am not seeing a web app server. A web app server is just part of the SOA platform. It is a run time for running certain Java containers but there is an entire SOA management platform that brings together development, operations management, optimization, policy systems on top of that. Probably something a fairly big company would need to do, and it would be a set of roll ups -- and I think that is going to happen around the end of the decade. The next thing is the world needs an open source service provider, global technology services, training, and support that could be a really big business and would support, by the way, the entire stack, including all of the key Linux with its players, and potentially Windows as well. I think that would be a catalyst for making open source real and it would be a really big play. In the end, they are providing it directly, and as a brand of service to others could make this market really take off.
Bill Coleman founded and was the first chairman and CEO of BEA Systems. Under his leadership, BEA became the fastest software firm ever to exceed $1 billion in annual revenue. As vice president of system software at Sun Microsystems, Coleman's team transformed SunOS into the commercially successful Solaris operating system. While at Sun he also founded Sun's Professional Services Division and co-founded Sun's Federal Division. With more than 30 years of high-technology experience, Coleman leads a seasoned executive team averaging nearly twenty five years of experience each.
|
|
 |
|
|
