Question and Answer with Hyperion CTO John Kopcke on Service Oriented Architecture, Web Services and Marketing Strategy
As more pundits promote the power of SOA, TDWI asks Hyperion’s chief technical officer how his company is riding the wave.
- By Eric Kavanagh
- November 9, 2005
Some say SOA is just the latest buzz word, but more vendors and consultants seem to be coming to grips with the revolution in application development, deployment, pricing and production wrought by this growing enterprise protocol. We asked Hyperion’s CTO, John Kopcke, how SOA fits into their strategy, and what impact it’s having on the bottom line.
As a platform player—and we presume that’s a classification Hyperion embraces—it would seem apropos that you push an SOA-enabled dashboard and portal solution as your flagship product. Is that the long-term strategy? Own the point of control, the actual interface used by executives to manage the business, top-down?
It’s best to step back for a moment, to agree on definitions. The first one is SOA. Of course, SOA as a concept is not new. Some people trace back to the 80s; I trace and give credence to it in the 90s with object-oriented technology and CORBA. What’s brought SOA to the top these days is Web Services. The idea of having coarsely grained modules that provide a service is not new; what was different with CORBA was stitching them together. Web Services make that much easier to do.
From an internal architecture position, Hyperion first started delivering SOA-based products in the 90s, [including] our acquisition of Brio; and our [financial] applications have been based on an SOA architecture from their inception since about 2002. [Hyperion] System 9, which we just introduced, highlights and extends the SOA architecture even further, in that we have Web Services interfaces with our dashboard technology today; and our dashboard technology is capable of participating within the industry portals that are out there. So, no, it’s not a long-term strategy; it’s something we’re doing already.
The way I’d describe the meat of the matter is that we’re moving from an application-centric architecture to more of a process-centric architecture. What are some of the vehicles for enabling this transformation? Technologies like portals, workflow engines; technologies like Web services and SOA; and Hyperion fully embraces all of that and sees us being well ahead of the competition with release of System 9.
A mid-sized Southern bank says it’s now using Web services for 20 percent of its business intelligence needs. They say in this coming year, that will begin to affect their BI tool licensing. Have you started to see any degradation of license renewal revenue as a result of Web Services?
Absolutely not. I’m not sure what capabilities they were previously using a BI tool to do, which is being replaced. If anything, the more an organization moves to an SOA, the greater value and greater need they have for what Hyperion does. Access to information and processes becomes much simpler; but I haven’t seen where an SOA architecture, in and of itself, removes the need for the query and reporting and analytics, and what I typically associate with BI. They may associate some other tasks. Maybe Cognos DecisionStream capability, EII/ ETL activities, as part of BI—SOA could certainly erode that. If anything, SOA is an enabler of what Hyperion does instead of a detractor.
Do you and will you work with smaller Web Service providers on a pay-as-clients-use basis, such that you’ll essentially “open” the Hyperion platform via SOA to specific, best-of-breed Web Service providers?
Again, SOA is an important part but is not the whole picture about how you deliver software as a service; it’s one vehicle for that, or for on-demand pricing. There are other technical issues with tracking that usage and reporting and protecting privacy, and how you’d price that. Certainly, if your technology is based on SOA, it’s easier for you to address those things. But there are technological inhibitors and business models that are the bigger challenges to get to that kind of pricing.
In the time-sharing days, when people would dial in to your computer, back when I was at Comshare and we were working with computer units; the pricing was computed on a mystical, magical formula, based on your connect time, and how much CPU you burned, what day of week it was, whether it was peak time, off time—customers paid that way. As we look at on-demand, those issues come back up. The difference is that software in this case is a service on your computer delivered via the Internet; but there’s still a complex set of software to manage all that, and it becomes more difficult because you have lots of detail you must track form a usage perspective. Organizations don’t like you reaching in and pulling out that kind of data. It gets CIOs kind of twitchy as you go in to read the meters, so to speak.
Do you see any degradation of business for Hyperion or in the broader industry, now or in the near future, resulting from SOA for enterprise integration efforts? Data integration? How far along the way is SOA to becoming a sort of de facto, built-in, integration solution?
From a Hyperion perspective, we had a record financial year which ended last June both in revenue and profits; and just had the best first quarter in the company’s history. So there are no indicators of an erosion from SOA in our business; and I haven’t seen any. From an overall software industry perspective, there are some parts of our industry that will be challenged by SOA, and those are companies in the data movement business; and those trying to sort out the complexity of the infrastructure; so that might hit ETL vendors, the EII and EAI vendors; but Hyperion doesn’t really participate in any of those markets.
As CTO, have you begun to reshape your development environment to reflect a stronger focus on Web Services? Have you created or do you plan to create specific teams dedicated to achieving a greater SOA-friendly architecture for the Hyperion platform? Do you have any staff dedicated to SOA research?
In earnest across the enterprise, we started that four years ago. I’d describe us as being in mid-stride with that. It’s not something we’re starting now; we’re ahead of game. In System 9, we not only have all our BI capabilities, but all financial applications combined as well. The way we’ve organized our development is we have a Common Technologies Group, and an Architecture Group who make sure all functional areas in System 9 conform to the architecture; and the architecture is all SOA in nature. We’re continuing to pursue best practices there; as opposed to having a research group exploring new best practices.
More companies are starting to employ an Enterprise Service Bus. Is this an application space that Hyperion will explore?
No, we’re not going to be doing that; but we’ll be leveraging it. This leads to one of my favorite topics, which dovetails nicely with all we’re doing. Four years ago, recent history, people would ask do you run Unix, Linux, Windows; that won’t be question moving forward. The new question will be do you run SAP, Oracle, IBM, Microsoft? They’re further abstracting the layer that companies like Hyperion look at, further away from operating systems. We still have machine language, assembler language; but I haven’t heard those terms in 10 years. Operating systems will go through same abstraction; and the reason is, this ESB will become fundamental capability of those infrastructures. WebSphere is probably one of the most capable infrastructure today; and what it means to Hyperion, is whether it’s SAP, Oracle, etc., if you don’t have an SOA it’ll be hard for you. What’s driving this is that all this nTier stuff we’ve been doing has been wonderful from a scalability perspective; but the complexity of it has become enormous; and this ESB is a very good technique to simplify that and make it much more manageable.
When I asked you a question about Essbase back in New Orleans at Hyperion Solutions 2005, you quickly diverted attention to the broader Hyperion solution set, noting that companies don’t buy specific products or technologies; they buy solutions. Then, in your System 9 release, there was nary a mention of Essbase, despite a recent quarterly earnings statement that gave one positive line of comment about Essbase7X sales performance. Is this the standard software vendor strategy of trying to shake an old image? Is Hyperion really trying to reshape and essentially expand its image beyond Essbase, hence the reticence? Is this careful wording around Essbase a product roadmap issue, or just a PR/marketing strategy?
So, the answer to that question: we’re not very concerned that Hyperion is so equated with Essbase that people don’t see more of what we do. What’s really driving this is a desire for simplification. When you talk about Hyperion, and what it can do, you can get bogged down in tons of product names. I sat once in a user group meeting and counted the distinct number of product names that customers had to know and understand, and it was 27. And the number of questions asked was vast: was it Analyzer, Planning; which product did what? So, what we’ve done is follow the lead of others: it’s Oracle 11i, 10G, My SAP; so we’re trying to go to a branding which is System 9: multidimensional analytics; enterprise planning; and all the other areas of functionality.
We recognized that Essbase is a very strong brand recognition; and as we pointed out, 7X revenue has been trending up whereas it had been flat or down the last few years. But the emphasis for not specifically pushing that brand name more and more is just to get to a simpler dialogue about what Hyperion offers.
What do you think about Oracle acquiring PeopleSoft, then Siebel? Does that have any effect on Hyperion sales or strategy?
The PeopleSoft acquisition, from a Hyperion perspective, has been a net positive, because when you looked at companies that were refocusing themselves in our area, PeopleSoft was probably doing a fairly decent job; and with the Oracle acquisition, that took steam out of that. They’re now focused more on the classical Oracle strategy, which we’re very comfortable with.
Siebel is dramatically different than PeopleSoft, in that the PeopleSoft culture was a mellower culture than the Siebel. So it’ll be interesting to see the fallout of two aggressive cultures coming together. But Oracle has been a very interesting competition partner of Hyperion for a number of years. As with many large companies, that are parts you compete bitterly with, and parts that you aggressively support. Hyperion is a big provider to Oracle and Siebel customers; hence now an even bigger provider.
Anything else you’d like to add?
Kopcke: Just one other point: I don’t think we have as unique a vision of BI and BPM as we once had -- Especially now that BusinessObjects has bought a planning company. We always intended that BI was a subset of BPM. Perhaps you noticed that Howard Dresner has joined us as our Chief Strategy Officer.
We noticed!
Howard has a great sound bite where he says that BPM is BI with a purpose. And I really like that, because when you step back and ask a business user about BI, they usually have no idea what you’re talking about. But if you go to IT and say BI, they’ll say: “Oh, yes: data warehousing, data mining, query and reporting—OLAP.” And after you walk away you realize that all you did was talk tech. If you go to business and talk BPM, things like event-based planning, customer profitability, activity-based costing… and you won’t have talked technology at all. It’s BI with a purpose. So we think now that we were the first to jump in with that point of view; and Cognos has followed, BusinessObjects has followed. So it’s all-encompassing, BPM.
Another way it can be characterized is that BI is focused on insight: How do we get more knowledge? Perhaps you’ve seen the commercial with two consultants sitting across from a business executive, and he says, “Let’s do it.” Then they say, “We don’t actually do it.” So, if all we do is deliver insight, we’ve only done half the equation. How do you drive performance? It’s insight and performance that completes the picture.