GCJ: Can you give us the history overview of VCompute, who you are, what you do, funding, employees, customers?

Hawes: OK, well, we started back in 2000, in September, moved into a secure bunker and we rolled out our initial infrastructure of about 200 nodes of equipment. It was real slow for a long time. I believe that the market wasn't quite ready for grid computing back then. I think it was being talked about, utility computing was something that people had mentioned and wrote some articles about, but people hadn't really adapted to it yet. I think a lot of it had to do with just understanding how to deal with vendors. There were security concerns and things that people had to settle in their minds. So it was like that for about a year, pretty slow. We had some business from the energy industry and some doing Internet intelligence-type work, people that wanted to mine the Internet for specific data. And then, I'm not sure what happened to the industry, but about October the year before last, it was almost like a light switch. All of a sudden, we were just bombarded with requirements for work on our systems.

Now, that being said, we have our own infrastructure, and we have hundreds of CPUs that we sell time on, but the energy industry many times needs thousands of CPUs, and they need them for very short-term durations. That's when we started looking around at the major vendors who had infrastructure available, and they weren't having a lot of success selling time on it either back then. We started buying compute time for the major vendors at a very good price and reselling it to these energy clients that needed thousands of CPUs. That's when you saw one of the press releases from Sun, I think October of last year.

So we have a policy that we're not going to roll out thousands of nodes for very short-term work. Typically, on our infrastructure, we book that infrastructure for very long-term type contracts, and that's typically the way it's been. Our entire infrastructure has been at 100% utilization since we acquired the cluster segments. We had a few gaps here and there, but not much. Now it seems we have actually become a clearinghouse for cycles. We work with many of the major vendors that have these HPC systems set up, to help them book time. So it's been a very successful business model, and it helped us quite a bit working with the major vendors to eliminate a lot of the risk associated with standing up huge amounts of infrastructure without having long-term contracts behind it.

So we've done very well in that aspect. We have partnerships with Sun Microsystems, IBM Corporation, and HP that supply us compute time. If they have it available, we'll buy it from them and deliver it to the clients that need it. We manage all aspects of delivery. So that's the business model.

Now, out of all of the industries, seismic is the hardest to work with, the energy industry, like I mentioned, need thousands of CPUs for very short-term durations, and they never really know when they're going to need it. You have to have infrastructure available for these guys when they come knocking at your door, and that's the biggest challenge in the energy industry. Now, in the other industries that we service like data mining or Internet intelligence and bioinformatics and life sciences, those are clients that typically will buy long-term segments of compute time from us. So it's been a very, very good model for us, because we can buy infrastructure and plan on it being busy for quite some time. We've had some very big wins and success over the last several years in these markets.

We have - some of our own equipment which is Dell, and early on we bought some Rack-Saver, I guess it would be Verari Equipment now. It's been an interesting business. The energy industry has, of course, been impacted by the high oil prices, so we've seen a lot of increase in demand there, a lot more exploration is going on. These companies are shooting more surveys, and we're rendering those surveys for them, so they know where to drill.

GCJ: On the Sun Grid purchase, I just wanted to clarify. I guess it's not only Sun, but also other places like IBM where you buy it. So for all of that computing power you're purchasing and then reselling, did I understand correct you're at 100% of all of that? So you're using everything or reselling everything you buy.

Hawes: Right. We haven't had an overabundance of cycles that we've purchased and not been able to sell. So typically when we go to the majors, we already have a project in hand and we buy X amount of compute cycles to satisfy those requirements.

GCJ: And I know on the press release a year ago from Sun, you guys were the largest deployment. Today is that still true, do you know, with them, or...

Hawes: I don't know that to be the case. I think from the energy sector it probably still is. I think Sun has changed their approach to utility computing. They have decreased the size of their commercial grid and rolled most of their infrastructure into that retail offering, at a dollar a CPU hour. And that's a completely different model. I'm sure you've seen the press releases on that. It's - you go in with your credit card, you buy a certain amount of time on a number of nodes. They supply you with access via the Internet, and you have to move your jobs over and that kind of thing. But that's somewhat limiting. We can't use the retail grid that way, because all of our clients require a very special setup. Many times Solaris is not used. It's different flavors of Linux. And it requires special configurations that don't really fit the model of the Sun retail grid.

GCJ: You alluded that you have some equipment that's yours that you sell or partial out, and then some that you've repurposed from IBM, Sun. Do you have any indication, or do you share the breakdown on how much is your equipment versus others, or do you not do that?

Hawes: No, we typically don't release how much infrastructure we have in the press. There's a number of reasons for that. One is we're watched very closely by a lot of the majors and people that are interested in competing with us. However, what we've found is a lot of people seem to think that they can set up infrastructure and compete with the majors and companies like mine by providing computing resources. But, so far, every one of them that's tried has failed. They go out and spend millions of dollars and set up a bunch of gear and then can't sell it. A good example of that was APPRO. They went out and, I don't know, spent millions of dollars, set up a thousand CPUs here in town, and couldn't sell any time on it.

Now, that being said, I think it's important to recognize that the utility computing industry has matured to some degree, but it is not yet a consumer-based service. Many of the clients that we do business with are clients that we've had relationships with for many years, and there's an element of trust that's required for these large organizations to put their production runs in your hands. These are critical, in many cases, critical runs for big oil companies. They have to be done by a certain time. They have to be done correctly. It impacts the ability for these oil companies to drill in the correct locations, and meet certain timeframes. So it's not like you can just run out and buy compute time anywhere and get your job done. It requires some detailed understanding of specific industries and how to approach those things.

The utility computing business is a very challenging business, and I think that's why many of the majors have had difficulty succeeding in that arena, because it requires some very detailed understanding of how to work with seismic and energy companies to supply those resources. You have to have a very strong level of trust in order to secure those types of agreements. That's where Virtual Compute has been very successful. That's why we have those relationships with the major vendors, because we are one of the few companies that have the ability to work within the energy industry and secure that type of work.

GCJ: Now, from a growth perspective for you, what - are you looking at other vertical markets to expand into?

Hawes: Well, yeah. We've found that the other markets are also very rewarding for us. - Life Sciences and Internet intelligence, I refer to it as Internet intelligence, but it's really data mining on the Internet. That type of business, people that have search engines, people that do specific research on data on the Internet, those organizations have had very good success with utility computing. Life sciences is where we really do a lot of our business. Energy is a big user of ours- because we're in Houston, we have a lot of the energy clients, but that is not our mainstay. A lot of people think it is, but it isn't.

GCJ: From a customer perspective, are there any customers in particular you'd like to talk to or any case studies you'd like to go over briefly?

Hawes: Paradigm was a very good client, still is a very good client of ours. The particular runs with Sun were successful. There was, I guess, a lot of difficulty initially getting them configured and running because some of the applications have to be adapted to the infrastructure. So there were some challenges initially with them, but as far as case studies, it's difficult for me to discuss my clients because many of them prefer to remain anonymous.

Paradigm was kind of unique because we were able to talk pretty openly about our work with them in the media, and that was something that was agreed upon with the client early on. But most of the companies in the life sciences area, people that are doing data mining, we maintain a very high level of confidentiality because their business models are so unique. That type of mining, and the type of data that they're extracting from the Internet, they don't want people to know what they're doing.

GCJ: Sure. Let me ask you then from a market penetration and market appreciation perspective, you've used the term utility computing. There's been a recent spat of press around virtualization, grid computing, SOA, and utility computing and how they all intertwine, or maybe they don't, or one leads to the other. What are your customers telling you? How are they progressing and what kind of tools are they using?

Hawes: Well, I think a couple of things that people need to recognize is that the fact that Virtual Compute Corporation has so many CPUs available - we have hundreds and hundreds of CPUs that we provide. We are in a very unique position to accommodate organizations that need business continuity or virtualization for disaster recovery and things of that nature. The fact that we store images of these client environments for their geophysical type of production, it's a logical extension for these organizations to protect themselves from a disaster standpoint on that same infrastructure, because we're familiar with their environment. Disaster recovery and business continuity, even after the hurricanes down here in Houston last year, we still haven't seen the organizations really take disaster recovery and business continuity seriously. They back up their data still and they store tapes and things of that nature. But we have the ability, and we have done this for several clients, we have the ability to store their image of their environments in our infrastructure. We have the ability to back up their data in real time and keep pretty close copies of real-time data in our environment. So if a disaster were to hit, they could use our infrastructure to support them during that period of time.

And the other point is that many of these companies that have their own clusters have not had organizations like Virtual Compute available to them in case of a disaster. So there were very few companies that had enough computing resource to match their environments. Clients are having difficult times getting the insurance companies to give them business continuity coverage insurance, or if they did it was very high, because they had no hot sites available, no virtualization. And that was purely because there weren't enough organizations that had that kind of infrastructure available to them. So that has been a point of focus for us because we have the ability to vacate production jobs and allow organizations to use our infrastructure for emergencies. Local and state, federal governments have the ability to come in and work with us. But we haven't seen the commercial industries really embrace the business continuity and virtualization yet. And I don't really understand that, to be honest with you. You would think that after all of the disasters we've seen recently that they would take that more seriously. But they don't.

GCJ: Is that an opportunity for you moving forward? Is that something you're looking at developing solutions around?

Hawes: Oh, absolutely. As a matter of fact, we already have a suite of solutions for that very situation. But people - when you start talking about business continuity and virtualization, all of a sudden you have to open up that box of issues that are associated with that. And when you open that box, you uncover a lot of weaknesses in these environments, and a lot of these organizations don't want to - I don't want to say they don't want to admit to those weaknesses, but they don't want to open that box. They feel comfortable, they've got tape backups, and they can buy additional gear if something happens. But I think you and I both know, when a disaster hits, in the time it takes you to get back up is huge losses. And in some cases, if you don't recover quickly, it has serious impact on the corporation.

So that being said, I think we'll see more and more of that. I think we'll see more demand for that. And because we have dealt with so many of the energy companies, we have a lot of these corporate environments' images stored and available. So if we had to blast out images on an environment for them to run through a disaster, we could do that very quickly.

GCJ: From a software tools perspective, what tools are you using and providing to your customers, and what are your customers using internally for the grid?

Hawes: We give them secure shell access to our infrastructure to the head nodes of the cluster segments that are provided to them. We provide tools such as Ganglia and others that allow them to look at the environment and see what's going on, on an almost real-time basis. The management nodes that we provide them have the ability for them to blast out new images or store their data. There are not any real fancy tools that we have gone out and purchased from major vendors. Most of the stuff that we use, and most of the clients use, is open source type of tool sets, such as Ganglia. And Ganglia's one of our primary tools that gives us the ability to do monitoring and management of the nodes remotely for the clients. So we typically run Linux operating systems, and so some of the tool sets that come with the OS are also utilized. But we haven't gone out and bought anything really special for our environments. Most of the open source stuff works great.

GCJ: What's next for the remainder of the year, or any big announcements that may be coming up that we can keep a lookout for?

Hawes: Well, we're doing a lot of business with IBM right now. I think that that's going to continue to increase in volume. Sun is, of course, going through a transition period right now, with Jonathan taking over and McNeely leaving. The downsizing of the company, I don't know what the future holds for their utility computing segment. We will continue to work with them. We provide a full spectrum of capabilities from the old 32-bit stuff up to the dual cores and quads that are coming out, so we're always pretty current. But we do use the majors like IBM and Sun to supplement.

Now, as far as our growth, we've seen huge growth over the last six months. Our sales have increased 2000% since October of last year, and that trend is continuing. We are really, really doing well in terms of revenue and profitability. And I think, out of all the companies that are doing utility computing, we, surprisingly enough, are one of the few that is actually making good money at this. Others are taking huge losses because they're trying to recover the cost on their infrastructure. But we didn't approach the market that way. We bought infrastructure as contracts became available, and we booked that infrastructure over long term, and we've supplemented the short-term requirements with outside vendors like IBM and Sun. We haven't really had to take the huge amounts of risk that the majors have. So we anticipate continued growth, continued profitability, with increased acceleration in terms of revenue.

close window