|Home - CEO Spotlight - May 03 Issue
CEO Spotlight: Peter Lee, DataSynapse
By Angel Mehta, Managing Director, Sterling-Hoffman Management Consultants
Angel Mehta: There’s been a lot of buzz about grid computing in certain pockets of the community for a while, but it still doesn’t seem like a mainstream technology. Talk about DataSynapse in the early days…where did the people come from? Where did the technology come from? What inspired the company?
Peter Lee: I think the guy that needs to get the key credit here is my partner and the company founder, Jamie Bernardin. The two of us grew up together …unusual, I think, in start-ups that the two guys who started it go back over 30 years. That’s a hell of a lot of history to have….
Anyway, Jamie is an ex-Wall Street tech guy. He has a Ph.D. with an Engineering Physics background so you know what I mean…he spent about 8 or 9 years on the street at different firms constantly building a variety of computing and data-intensive applications and it was really through his experience hands-on in building these mission critical apps and constantly reinventing the wheel in terms of trying to solve those bottlenecks that he began to get a strong idea that a general application platform would be of tremendous value.
From my side…I mean, I came out of JP Morgan – and in both the capital markets and the investment banking side , the user case for Wall Street in terms of solving these important bottlenecks in the front and mid-office has an enormous value impact on the business. For example, if you’re a trader and you don’t get your risk reports in the morning before you open up for trading… you have one hell of a problem on your hands. The bottom line is, the relevance of the technology to the financial vertical resonated with both of us. It’s a very generalizable technology…but I think the market focus has really helped anchor the company through these very turbulent times.
Angel Mehta: Am I wrong in saying that the technology is still ‘pre-chasm’, per se? I just read an article in Scientific American on the possibilities of Grid Computing and it sounded to me like it’s still very much an emerging technology…but based on some of the traction you folks have…what stage of adoption would you say we’re at?
Peter Lee: I think it’s very dependent on the customer vertical. We really cover three industries right now: finance, energy and the government sector. In the finance sector, by way of example, whether you call it ‘grid computing’ or ‘distributed computing’, this type of technology has been in use for well over 10 years. There are extensive home-grown systems very tightly coupled to their underlying applications and constantly reinvented for each specific business unit. You know, any time you have a customer that has been immersed in your technology field for years and years, it’s a difficult sell. What we bring to the table to differentiate is a solution that’s out of the box. It’s an infrastructure generalizable platform solution that can streamline all of their costs and get them out of the ratio of maintaining infrastructure and moving that ratio more heavily into innovating on business applications and business logic. Grid computing as it’s currently positioned in the media is not a ‘Science Fair” project-type undertaking. There’s too much of that going around these days.
Angel Mehta: Can you give me an example?
Peter Lee: Sure. We have a major energy & oil exploration client …where a very typical job process for them in the past was to run a seismic imaging application that will take a couple of weeks or more over hundreds of nodes. Using our software, they’ve accelerated that into an overnight process.
Now, when you start examining the impact…it’s not just the speed in terms of being able to extend a constrained architecture into an infinitely scalable architecture. What you have to really think about is the cost savings there are in terms of people and manual intervention and work around and failures. For example, with the oil company, it’s very frequently the case that bad data corrupts the run a week into it then you have to restart it. Now, with our software platform in a kind of business critical context you have failover and resilience so that when you have errors on the network we RECOVER from that.
Angel Mehta: You said grid computing has been around for a while…10 years, in the financial services vertical. So what makes it different today?
Peter Lee: You’ve got a few factors moving in our favor. I think first and foremost there is this extraordinary emphasis on doing more with less.
Angel Mehta: As in, do more with less hardware?
Peter Lee: People always think that way, but that’s not all there is to it. Much more importantly is scarce people resources and scarce time. So what you’re talking about is delivering increasing business functionality with lower dollars, with lower absolute infrastructure costs with fewer people. I think what grid computing does is it takes a lot of existing IT paradigms such as: “I need to develop an application, I need to couple it tightly to an underlying infrastructure platform and then every time that infrastructure platform evolves I go through a huge and burdensome testing, QA and redevelopment process. At the same time, every time I want to roll out new business functionality I have this enormous production trade-off between stability and new enhancements.”
That’s the current IT paradigm: a bunch of different silos with extraordinary latent capacity and over-provisioning. Grid computing flips it on its head. It says, “Why don’t we take a more holistic look at the total workload of the enterprise and the total resources of the enterprise, people and hardware, and budget and why don’t we reconfigure this to have much more of a utility computing or an on-demand computing model, where when one group could use resources that the other group doesn’t need we can share among those groups”.
I think that process of breaking down existing application barriers and unit barriers is proceeding at a pace far more rapid than we had previously imagined.
Angel Mehta: I want to go back to the time you thought about the verticals you were going to go after.
Peter Lee: Sure.
Angel Mehta: It makes sense to me that you went after financial services right away, given that you were domain experts and that sector as a whole tends to have early adopters that are under incredible pressure to make the technology work….plus risk management would be low hanging fruit, clearly. So that was a no-brainer. But how did you pick energy and government?
Peter Lee: We attacked the seismic imaging...well, really the exploration and production energy vertical for a couple of reasons. First, there’s another clear value proposition that represents low hanging fruit.
In the exploration and production space the ability to find and make decisions about lucrative energy deposits wherever they may lie in the earth’s crust turns out to be a very important business function. Right now that business decision-making process is in a lot of pain. Let’s say you’re either buying or selling properties with another energy player or there’s a competitive bid offshore in some offshore track the government was leasing and you need to make a quick study and judgment about how many billions of dollars of energy deposit is underneath and how accessible it is. You cannot AFFORD to wait six months for your analysis to come back. The decision to drill or not to drill…I mean, if you hit a dry well or a dry hole that could be a very serious negative investment.
Angel Mehta: It could cost you millions to come up dry…
Peter Lee: Right. So there’s the recognition that the more analysis invested up front, the more likely it is that they can reduce risk before they proceed.
Angel Mehta: There’s been a lot of hype about the applications of grid computing in life sciences, but DataSynapse doesn’t seem to have targeted this market. Why not?
Peter Lee: We looked at that market, and you’re right there’s been a lot of talk about the human genome and using computers to crack a number of proteonomic sequencing and so forth.
What we found in our early level of client dialogue is that while there is a tremendous and insatiable requirement for ‘computer processing power’, we haven’t yet discovered the same type of value proposition as we have for other verticals. I don’t know whether that’s because the drug discovery process is such a prolonged multi-year, multi-phase process from start to production finish. Because in those cases, if you can only influence a small piece of that chain, you’re really not influencing the total end-to-end that much. It’s different than, say, in finance - where if a process takes a day, I can bring that into an intraday process. With Seismic imaging, if you’re in a week long or month long process - I can accelerate that overnight you have an immediate payback.
Angel Mehta: So if it’s a 10 year drug discovery process, with clinical trials, etc., you’re saying it’s just harder to map the value?
Peter Lee: Right. it’s just not clear. I think the value proposition hasn’t yet really fully gelled or has been articulated.
Angel Mehta: Let’s switch gears a little bit. When I think about you and Jamie, it makes me curious about the whole ‘doing business with friends’ issue. Has the business ever put a strain on the relationship?
Peter Lee: Sure. I mean, there are many times when we’ll disagree but I will tell you that when you have a 30-year history…I mean, we’ve known each other since we were little little kids. So there’s an enormous amount of trust and confidence that’s built-up and I think that’s the basis first of all.