Thursday, October 02, 2008

Interview: HP's John Santaferraro on Latest BI Modernization and Data Warehousing Strategies

Transcript of BriefingsDirect podcast recorded at the Oracle OpenWorld Conference in San Francisco the week of Sept. 22, 2008.

Listen to the podcast. Download the podcast. Find it on iTunes/iPod. Learn more. Sponsor: Hewlett-Packard.

Dana Gardner: Hi, this is Dana Gardner, principal analyst at Interarbor Solutions, and you're listening to a special BriefingsDirect podcast recorded at the Oracle OpenWorld conference in San Francisco. We are here the week of Sept. 22, 2008. This HP Live! Podcast is sponsored by Hewlett-Packard (HP) and distributed through the BriefingsDirect Network.

We welcome John Santaferraro, director of marketing for HP’s Business Intelligence (BI) portfolio. We're going to be talking about the intersection of BI in the context of not just business value and outcomes, but in the context of Oracle, a major data applications middleware and BI provider, and HP as prominent systems provider, as well as a prominent BI services provider.

We're going to try to figure out how this plays together. Then, we'll look toward the future of BI in the context of some major trends, such as service-oriented architecture (SOA), master data management (MDM), and bringing more automation to the delivery of intelligence from systems and data to those users who need it at the front lines of business. So I want to welcome John Santaferraro to the show.

John Santaferraro: Glad to be here, Dana. Thanks.

Gardner: First, let's set the stage and get a level-set about the Oracle-HP relationship vis-à-vis BI, because we're here at Oracle OpenWorld. Oracle is in the software side of things predominantly. You’ve got both systems and services. Perhaps you could paint a picture of how this fits together.

Santaferraro: It’s been a great and long relationship that we've had with Oracle since they were first building and releasing a database. We had folks in our labs that understood this idea of databases and data warehousing, and they were actually building and architecting our systems in a special way with things like massive I/O, massive memory to address -- the kinds of things you need in a data warehouse and query environment.

Back in those days, we were actually building our systems to handle data warehouse workloads, when everybody else was still focused only on the regular online transaction processing (OLTP) kinds of transactions in the enterprise resource planning (ERP) systems.

Because of that natural connection that we had with what was going in our labs, and what Oracle was doing, we have from the very start built a tight relationship with them from an engineering perspective and a good market perspective. Oracle is very clearly a leader in data warehousing and BI, and we augment that with the systems that we have developed to run in an optimized way with Oracle, as well as some other services that we bring to bear.

We recently bought a company called Knightsbridge, which was known as the go-to company for anybody who was doing data warehousing or BI and who ran into problems that nobody else could solve. Everybody knew that if you went to Knightsbridge, there were people there who could solve those problems. So it’s great to have them at the center of our global BI services organization. This company has taken their methodology and their expertise and has transferred it to folks around the world.

The other great thing about the acquisition of Knightsbridge is that they have real deep expertise in their various vertical markets -- health and life sciences, communications, financial services, retail manufacturing. Because of that, the Oracle-HP relationship is strengthening.

We are more than a systems provider and more than a services provider. We are delivering real solutions to our customers. We can come alongside of anybody, talk to them at the level of the business, and be able to build data warehousing and BI solutions that are mapped to the business, not just technology.

Gardner: I just got back from listening to Thomas Kurian at Oracle describe their full portfolio, and they’ve really put together quite a full lifecycle approach around the gathering, cleansing, and organizing of data, integrating it from disparate sources, managing the scale of huge loads, making this closer to a real-time value. They're also exporting middleware for application integration, creating the BI analytics, and then delivering that back out to those business applications.

It’s quite an impressive portfolio. They've been putting it together for quite some time, and they’re also quite proud of the metrics around the performance, and getting closer to that real-time nirvana. Tell us a little bit about how what Oracle has done from the lifecycle perspective and what you think are important aspects of the services’ side of making organizations readily able to let exploit those technologies.

Santaferraro: What you described is very much a product lifecycle in the data warehouse and BI space. Along with that, you can go in two directions. Along with the product lifecycle, there is actually a system lifecycle as well. Anytime anybody says to me that they can make data warehousing simple, I react, because the truth is that it’s very complex.

The processes you just described are extremely difficult for any company to work with and navigate through. Add to that the whole infrastructure piece of it. The more you move towards “operationalizing” BI, suddenly the more important the infrastructure becomes.

A lot of time we get calls from customers who are trying to deploy data warehousing solutions. They'll be in test and development and are supposed to perform, and they've got users out there who are expecting to click on a button and get all of the information back within a matter of seconds, and they can’t figure out how to make it work.

So they call the HP storage folks and they say, "Hey, we’ve got a storage problem. What’s going on here?" And, the storage folks say, "Well, wait a minute, it's not storage. That sounds like the database." So, they call Oracle, and Oracle says, “Well, that’s not us. It’s not the database. It must be a server problem.” So the customer has to go back to the server guy. We have people that will lose weeks of time in deploying their systems, because the entire lifecycle is extremely complex.

What we really do is look at how can we come alongside of Oracle in our labs and figure out how to build those systems with Oracle, pre-installed, pre-configured, and pre-tested, so that what the customer is getting is ready to go out of the box. It takes the guesswork out of all of this implementation and development that they’ve got to do.

I had one customer who lost a week in production, lost a week in test and dev, went into production and made the same exact little thing. They forgot to turn on a synchronous I/O on their storage system. It’s just a basic little problem, but it cost them another week in production time before they were up and running.

So, we’ve got solutions like HP BladeSystem for Oracle Optimized Warehouse. We have about 50 reference configurations that help take the guesswork out of deploying these.

Gardner: This is really more than just one hand washing the other. This is three hands washing each other. We have the systems integration and specialized software, which is created through products, integration, and technology innovation, and then the opportunity for that third hand of services to come in with methodologies and best practices, for preventing those gotchas.

Santaferraro: Exactly. And then, on the services’ side, here are people who have walked this path before. They’ve done it before. My recommendation to companies who are out there trying to do BI and data warehousing and are hitting difficulties is, “Why not go find somebody who has done it before?”

You really don’t have to do it alone. There are people out there who have walked this path. They’ve done it. They know the gotchas. They have accelerators. They have ways of making it all come together faster. And all of that translates into more business value. If I don’t have to spend as much time in deployment, as much time in all of the testing and trying to figure out what is wrong, then I can be investing my time and my effort in developing real business innovation and real business value.

Gardner: And, of course, in the field there are many different companies that are at different places on the path toward some of these goals. For those that are deeply into BI and recognize the value of getting this lifecycle, elevating the data, getting that good quality data out, and then be able to work with it, what’s the next step?

I’m hearing some buzzwords nowadays about operational BI and even BI modernization. Tell me little bit about what these mean, and are these in fact the next chapters in where companies will be taking this capability?

Santaferraro: Yes, these are definitely the next chapters, and you're seeing right now probably about five percent of companies out there -- the ones who are on the leading or bleeding edge -- already doing Operational BI and BI Modernization.

Operational BI has to do with this idea that I have all of this data in a single place, it’s accessible, and it’s fairly well cleaned. I don’t think anybody has perfectly clean data -- that doesn’t exit -- but once it’s there, what do I do with it?

We're finding that customers want to do two things. One, they want to get that information to everyone across the organization, as well as customers and partners, and they want it to be actionable. So how do I get actionable information in the hands of everyone across my organization who needs it?

The second thing I see is people wanting to do with operational BI is actually take the analytics that are driving their systems, and embed them in the business processes or in the business applications. When a loan comes in to be underwritten, you want to have the right rules that don't put you in a position as a bank where you end up with a bunch of loans that you can't sell in the secondary market, or going into default. Everybody is aware of that problem, right?

How do you take the analytics and discovery that you’ve made and put it right in the applications, so the decision is automatically made by the application or so somebody has it right there. As they are using the business application, they have the information to make the decision right there at their disposal.

Gardner: And is that what you call operational BI?

Santaferraro: Yes.

Gardner: Now, this also raises in my mind a question about the capabilities that a services oriented architecture (SOA) offers -- governance, bringing services like BI as a service into play with applications, but at the right point in time. So it's exercising governance policy; learning from your mistakes, and building on them. How does what you’re describing as operational BI and SOA fit together?

Santaferraro: It’s a great question, because when I hear people talking about SOA, I primarily hear them talking about business services. How do I take these mammoth applications that I’ve built, reduce them into reusable business services, and be able to use them effectively across the organization, instead of replicating them all over. The real opportunity comes when you have these business services in operation and you begin to bring in information services as well. Take customer profitability, for example. That's not really a business service. It’s an information service.

A lot of analysis has to go into the mix for companies to figure out or answer the question, "Who are my most profitable customers?" If you can figure that out, and give every customer a rating, then that information service again becomes a service within a SOA that you can actually use and distribute in a very useful way all across the organization. You can send it to the call center, send it to the sales force, send it to the Web, and send it to the ATM transactions that are happening. So there's a whole opportunity of information services as a part of SOA that haven't even begun to be tapped.

Gardner: It’s sort of the intelligent implementation of BI as a service?

Santaferraro: Absolutely.

Gardner: How does that differ from BI modernization?

Santaferraro: Modernization is built around this whole concept that folks started doing data warehousing 15 to 20 years ago. It’s a fairly old technology; yet it’s still very useful. It’s still something that companies need to do, but a lot of new technology has come in and new kinds of data. We are discovering that data warehousing had great value. It has all the information in a single place. It made information accessible. You could now do analysis.

Gardner: But it was largely structured data.

Santaferraro: Exactly. Now we have other kinds of data coming. What about email? What about document management systems, and all the documents that are being digitized? What about new types of data like RFID? What about GPS data? There are all these new types of data, and we're discovering now that the data warehouse bubbled up.

It's a great value for BI, but not everything has to go into the data warehouse. In fact, we’ve discovered with a lot of our customers that as soon as the data warehouse gets to a terabyte, about 70 percent of the data in that data warehouse never even gets touched or used.

So companies are spending enormous amounts of money to build these massive data warehouses, and a lot of that is not being used. Modernization is about figuring out what data needs to go into the data warehouse and what needs to be delivered through the enterprise service bus (ESB). Are there certain things where you can just embed analytics out at the application layer and do the analytics out there? Are there other types of data that should be just cataloged at the user level?

Gardner: Metadata, for example?

Santaferraro: Yes, and metadata becomes the rich side of definitions around that content, that actually brings it all together for the sake of the user.

Gardner: Regardless of where it resides?

Santaferraro: Exactly, and that becomes active metadata by the way. It’s no longer just this metadata that sits below for the data folks to understand what’s there. It’s active metadata that the users are using to understand the information that they're looking at.

Gardner: I suppose that, over time, that’s going to also include events?

Santaferraro: Absolutely, events and then tie right into the new complex event processing (CEP) systems. One of the opportunities that I’ve not seen tapped into by any software companies is this whole new world of information delivery.

So, if you’re operationalizing BI, if you’ve got a modernized BI infrastructure with data provisioning in place, and it’s not just the data warehouse -- you’re basically trying to get it out to all these users across the enterprise and embed it in business processes. There needs to be the design of a brand new information-delivery system that actually can handle all of these kinds of data to the desktop, to the application, to the hand-held device, or wherever it happens.

Gardner: Without belaboring this point, what sort of technologies are you looking at? Is this syndication, publish-and-subscribe, terminal services? What do you use to get that out there?

Santaferraro: I would say, yes. Because, as I said, I haven’t seen anybody that’s done it yet.

Gardner: Good, a big opportunity there. Okay. We've talked about this modernization of BI. This is happening in the context of other trends, of course, for virtualizing our data centers, and a lot has been done to virtualize storage and data over time.

We're going to be bringing in more kinds of content. We might even be getting content and services off of clouds, other people's public services or perhaps a cooperative private federation among business partners, all of which has to be managed and accurately projected back into the application services and processes that people use. It sounds very interesting, and is a much easier sale to the C-class, the corner office in the organization, because this really helps them in the way they do business.

What can companies do in terms of exploiting these technologies, getting those business outcomes, and, I suppose most importantly, how do they get started? As you say, this is not trivial. It’s complex and needs to be done properly.

Santaferraro: Most companies are started right now in BI and data warehousing. What I hear a lot of customers say is that they either are not getting the value out the investment they are putting into it, or they don’t know if they are. So I think it really makes sense to kind of pause where you're at and bring in some experts to do an assessment.

We do a lot of work with customers. We look at the vision, the strategy, and the planning behind data warehousing and BI, and because of our depth of experience, we can come alongside our customers and help them figure out what’s working and what’s not to put value on where to really invest moving forward, and help drive that forward in an intelligent way. Why not do BI with some intelligence behind it?

That’s one thing. The second thing is that with operational BI on the horizon, we’ve got a lot of folks within our organization who understand the potential of what could be done with BI in a bank? What if you could have customer profitability, customer segmentation services, and offer optimization at every point of sale? So, for the teller, for the ATM service, for the call center, wherever somebody is interacting with a bank, all of that information is right there with them.

What we find is that people have been so caught in the world of reporting and just basic analytics and online analytical processing that takes place in the back room. We think that it also makes sense to move to this next level. Bring in some folks who understand operational BI and let’s dream together and figure out if you could actually have these capabilities, what could you do with your company? How could you transform your relationship with your customers and your suppliers?

It's basic vision strategy and planning, too. Let’s get together and dream about operational BI, and figure out what your company could become? We actually believe that in the next five to seven years that there is going to be a major restructuring of leaders in every single industry. The ones who come out on top are going to be those companies that figure out how to use BI to transform themselves into competitive leaders.

We want to be there with our customers to make that happen for them.

Gardner: And this is not just for them to actually find new markets, but to uncover risks that they wouldn’t have been able to uncover until it was too late. And we’ve seen examples of that -- and perhaps to focus on what the right businesses are to be in and not to be in? So it’s not just how to make things better, it’s also risk mitigation on what to avoid?

Santaferraro: Absolutely.

Gardner: Very good. We’ve been talking about BI and some of the next chapters in BI, particularly in a context of a longstanding partnership between Oracle and HP. We’ve been joined by John Santaferraro, director of marketing for HP’s BI portfolio. Thanks very much, John.

Santaferraro: Thanks a lot, Dana.

Gardner: Our conversation comes to you today through a sponsored HP Live! Podcast from the Oracle OpenWorld conference in San Francisco. Look for other podcasts from this HP Live! event series at hp.com, as well as via the BriefingsDirect Network.

I'd like to thank our producers on today's show, Fred Bals and Kate Whalen. I'm Dana Gardner, principal analyst at Interarbor Solutions. Thanks for listening, and come back next time for more in-depth podcasts on enterprise IT topics and solutions. Bye for now.

Listen to the podcast. Download the podcast. Find it on iTunes/iPod. Learn more. Sponsor: Hewlett-Packard.

Transcript of BriefingsDirect podcast recorded at the Oracle OpenWorld Conference in San Francisco. Copyright Interarbor Solutions, LLC, 2005-2008. All rights reserved.