Analytics gains respect
Gallant: I want to spend a little bit of time looking at the different markets you're in and what the differentiators are. If you take the analytics market, how would you define Sybase's key differentiator? What makes you different from SAS, Oracle, IBM, any of these players that are in the analytics BI market?
Chen: Price-performance. And the reason it is price performance is because of a patented technology. This again is one thing that I cannot claim credit for. When I started, we had already bought a company called Expressway two years before I came. They have a column-based architecture. What a column-based architecture is good for is to be able to sort out...massive amounts of data and to look for certain relationships without having to create a lot of temporary disk storage.
Now, for many, many years - until the last three years - people have been fixated on generating records. Analyzing the relationships between records has not been top of mind. The people who really need that are the Toys'R'Us and Walmarts of the world. They're doing pricing models and SKU tracking and all that. They buy big from Teradata.
But that's very bleeding edge. We have a departmental solution that is much, much cheaper and much, much, much faster than anybody else. Slowly but surely, we have, like, 2,000 customers around the world.
The first breakthrough was actually the Patriot Act. That brought awareness to the banks. They said, "I've got to find a way to do this more real-time. I can't run it as a batch every day at 4:14. This has got to be done in about ten minutes or so." And so, that brings the first wave on.
Then, gradually, as the market gets more sophisticated, it gets into all kinds of things, like pricing. People who want to sort large amount of data. Billing, for example. Our architecture is good for it. It requires less storage it's much faster - near real-time.
Knorr: Do you think there's a revolution going on now in the analytics market with columnar and NoSQL databases - particularly putting that stuff on the cloud, with all that extensibility and scalability?
Chen: I don't know whether I would use the word "revolutionary." I think it's finally gained respect.
Knorr: Do you think there'll be a broader range of applications now? Because it's a relatively small market right now, right?
Chen: Yes. Absolutely.
Knorr: It's very also very specialized, because only a limited number of people really know how to use analytics to its advantage. Do you think there'll be broader user interfaces that are more suitable to business analysts, that kind of stuff?
Chen: Absolutely. Here's the reason why: What drove our success in the last couple of years, after all this compliance reporting and stuff, was risk management. Enterprise risk management is in the board room of every company right now. In fact, I think the new SEC rules require the board to be aware of enterprise risk.
Column-based architectures are extremely well suited for these kinds of applications. So in the last two years, we came up with something called the Risk Analytics Platform. We brought a bunch of traders in who understand how it works, and they helped us design this on top of our analytics engine.
And then, most recently, you probably have seen that we invested in something called complex event processing. That's to graph data, structured and unstructured, to bounce against these templates and the backend trading data, the account data, to create a risk profile on a near-real-time basis. So if your users had deployed that three years ago you could probably have avoided the counterparty risk issue that the market has seen.
So the answer to your question is: Yes, it gets into mainstream apps now because this fundamental architecture allows these large amounts of data to be analyzed, and analyzed at high speed. This is a new requirement: real-time analytics and real-time needs.
I was with some customers on Wall Street last weekend. It's very easy; everybody does risk management. I mean risk profiling has been done since the mainframe existed. So what's the big deal? I mean we all of a sudden sound so intelligent because there was a crisis and the blow-up and all that.
That's not it. There are two things that have changed. Number one, the priority of risk management has gone to the boardroom from the analytics department buried under the CFO. Number two, the real-time requirement and the reporting requirement [to send risk information to] the government is new - and rightfully so. We now realize how esoteric the instruments are in the financial world.
So you could take that example, and you could put it in health care, you could put it in securities, in businesses, in government. It could be in energy trading. It could be in grid distributions. We sold a lot of analytics servers to the China National Grid. And they use that to manage the grid. They don't have enough power, period. So somebody needed to manage the provisioning of that, and all this data that's coming in rapidly.
Next page: Is BI BS?