BEA Looks Ahead
Rob levy is executive vice president and chief technology officer of BEA Systems. He leads the companys efforts to align BEAs technology vision with its business strategy. Levy recently met with Senior Editor Darryl K. Taft at BEAs swank new San Francisco digs to discuss the companys technology future.
What are your plans for the BEA office of the CTO for the next year to two years?
The basic premise of the office of the CTO is to predict the future and translate that future into what we should be doing at BEA. We are the people with the crystal ball. But it really translates to multiple technology strategy areas.
One of the areas that the office of the CTO owns is global architecture issues. So we have a group called the corporate architecture group that reports to me, with Paul Patrick as the chief architect for BEA. So hes sort of the overseer of the horizontal view of all of the technology implementation issues for BEA.
Our mSA, microService Architecture [the architecture of the BEA SOA 360_ platform], falls under that premise. Sometimes its a technical issue, like BPEL [Business Process Execution Language] and how we implement BPEL across the product lines. Of course, some of the product lines were ours originally and some came from an acquisition, so we need to figure out who is doing what when, and how do we make that look the same now.
Then we have an emerging technologies and standards group. Those are the people who sit on the boards of Eclipse, OASIS, OMG [Object Management Group] and so on. Their job is first and foremost to influence. So we look at the standards body and we look at what we think is important for our customers, and where we want to be leaders versus just followers. Some standards we believe need to be pushed and some standards just need to be followed. Then wed decide if we want active participation. Then we bring it back to engineering as far ahead as we can to make changes to the physical product technology. This is always a tricky game, because engineering is always a 12- to 18-months-in-advance-notice thing so they can make physical changes to products.
Another function is the deputy CTO function or the evangelist function. These are the people that talk to the analysts, talk to the press, talk to the customers. Theyre mostly outbound facing. Up to 60 to 65 percent of their time is spent on talking to customers and people outside the company. They also are responsible for research subjects. So those are the guys who two years ago were researching Web 2.0 and how Web 2.0—which at the time looked like a consumer play—would affect the enterprise and what we should do about that. And these really are the people I look to to become the go-to guys or knowledge experts on areas that are truly technology trends and not just product trends.
Then I have an innovation and incubation team thats responsible for identifying new technologies and playing with them—put them in a lab, roll up our sleeves and see what happens. We might look at a function feature in one area and find that we can put that together with a feature from another product and then we can create a new product. And thats where new SKUs come about.
And the last thing we do is what I just did today. Theres a portion of the office of the CTO function that is very similar to what partners in VC [venture capital] firms do. We look at the future, they look at the future. They put their money behind technology bets; we put our money behind technology bets. I spend some significant time, kind of in bursts, talking to the VC community, keeping an eye on what theyre doing and on what technologies they think are important. And getting an opportunity to take a look at their portfolios, which is how we sometimes end up buying companies or partnering with companies.
So, where are you taking it?
I think whats going to become more and more important is the emphasis on innovation.
The next emphasis from the office of the CTO is really about creating innovation, not just through making things happen, but by formalizing the process of innovation.
So, for example, were putting together the plan for inside incubation formally. So, people in engineering can step away from an engineering day-to-day job for six to 12 months to be able to develop a prototype or [perform] proof of concept of a product. One of the main problems with innovation on the inside [of a product team] is you spend most of your time writing code for tomorrow, and your head of engineering says, “I need it yesterday.” That doesnt leave a lot of time for innovating or creating code for future products.
I also see a large consolidation happening in the marketplace right now. So when we predict trends, its easier to do when things are going smoothly. But with the consolidation happening in the market, I expect to see some major turns left or right because you consolidate two things that might not have been consolidated before.
Its fairly obvious when Oracle goes out and buys a PeopleSoft, which is an application, and then goes out and buys another application vendor. Its buying applications. But when somebody goes out and buys something thats not directly in line with what theyre doing, that kind of creates a more difficult task for us in the office of the CTO in terms of predicting.
I think one of the things thats going to be really interesting to look at is how Salesforce.com moves forward. And not just them, but it seems like there is a trend now in the secondary tier, the people that are starting new stuff. The Googles, the Salesforce.coms—young companies, lets put it that way. Its not the middleware platform, but its a platform. Salesforce released their own application platform. Google is building a platform for their applications, which are the widgets, and now they are building enterprise-level applications. And Yahoo is building Pipes. It makes it interesting to figure out how applications will be delivered and developed within the next two to three years.
That will create a need for a whole slew of software and services to support whatever the model is that emerges. And everybody has a slightly different view of what they think will emerge.
We need to figure out how to enable the new paradigm of building new composite apps, as well as how we continue to support the hundreds of thousands of customers we have that have what is loosely termed “Web applications.” This is really interesting in that I see Web applications that are now legacy applications. So, the first wave of Web applications—the B2Bs, the B2Cs—are now becoming legacy applications when you talk to CIOs. So legacy is not just mainframe applications anymore. But those legacy applications need to be able to talk to composite apps. So we need to do both.
What are some of the ways youre thinking of approaching these challenges?
There are a couple of things. The last time we talked, we talked about my concept of humans matter [in the IT equation]. This is one of the things that I think is going to become really important. This problem was always perceived as an efficiency problem. Things like extreme programming and data center management—all of these IT activities were really focused on how do we make the process more efficient? It has its phase. The next phase of growth for the industry will be the humans matter wave, which is how do we enable the free-form, free- flow behavior pattern of humans? And instead of objecting to it and pushing it out of the system, catering to it and bringing it into the system.
So, for example, what we think would be an important change in applications is not the application development itself; its the process of application development for a composite app and how you bring collaboration. How do you bring people in?
We just released a whole set of social computing tools for the enterprise that cater to the basic behavior patterns of things people want to do in free form, but inside the governance and constraints of IT organizations to make it work. So, instead of saying, “No you cant,” which is what IT always did with IM [instant messaging], were doing the opposite. Were saying, “Take it and bring it into the portal under a set of governance, under a set of security elements and under a set of policies that give enough control for IT but that allow users to do Web 2.0 things without breaking control of IT.”
So if you look at PEP, our Pages, Ensemble and Pathways products, those are three products that touch three areas of how you build mashed apps, how you build widgets, how you do smart tagging and how you do wikis, RSS and blogs, all inside the portal. And then how you put all these things together to create a mashed app.
We thought of that two years ago.
The next big things that are on our plate include BI [business intelligence] and visualization. Once you get the ability to pull things together and create composite apps … A year ago we predicted that the next step for composite apps would be the need for social computing and the other thing would be metadata about whats inside.
Last year we bought what we now call ALER [AquaLogic Enterprise Repository], which is a product that supports the metadata about whats going on. So if you look at whats coming out next, all of our products are connected to ALER, which is a very important piece for us. We said the next thing people would need would be governance, and then theyll need to start touching things they didnt touch before. When they do theyre going to come back and ask for more. Theyre going to ask for complex event processing. But if they do that theyre going to need in-memory caching and real-time buffer analysis, real-time BI, real-time visualization.
Page 2: BEA Looks Ahead
BEA Looks Ahead
You mean like a DataPower type of thing?
Yes. Well, thats interesting. We have a tendency in the industry to give something a name that has already been there and then we reinvent the name. For instance with BI, what we talk about today in BI is not exactly what we meant five years ago when we said BI.
For me the appliance world is like if you look at the problems we had when we moved from the mainframe to distributed computing. The first set of distributed computers were a computer with an operating system that was relatively generic, but ran a specialized set of software. The first things they did was data entry. That was an appliance. It was clunky and big, but it was an appliance. It was a specialized hardware/software combination to do a single business task. Now today all thats happened is that they have shrunk in size. So I look at appliances and Im thinking for the next wave of things to happen those people are looking to do things faster.
Just like when we moved to distributed computing we couldnt wait for the old transactions to go from the source through to the back end and back. That process was too long. So what we did was move parts of the software into the distributed world and created applications that you could put in a bank. Its the same platform that we call distributed, which I feel I can call appliances, and we added more stuff to do out at the edge. And we kind of separated the business process then from data entry that ends up in a batch process under CICS, to do data thats only local to you thats in real time or near real time … and then move it over to the batch.
I think whats going to happen next is exactly that with localized SOA [service-oriented architecture] appliances—with ESBs [Enterprise Service Buses] included.
Youre still talking about hardware here?
Yes. … Right now people look at appliances and theyre really customized hardware/software. I look at Linux and I think to myself that this is really interesting the way that Linux is slowly emerging as the generic appliance operating system. My Tivo runs on Linux. There are consumer appliances running on Linux. Why not have a hardware machine thats customized to be efficient in running Linux, but that runs a different version of the software than we did before? We said we cant afford to take WebLogic on Linux and take it and make it smaller and it becomes a mobile WebLogic.
I predict that appliances for SOA are going to become more important, possibly running on Linux. Not just customized XML transformation boxes, which also have a place. But those are for me more like Cisco [Systems] routers. A Cisco router is an appliance; its a piece of hardware and software that does nothing more than switching. They do program it, so it is a computer, but it does only one thing. Im talking about more of a generic box you would put at the edges to make service bus and data transformation faster. I could see localized in-memory databases that support real-time processing that runs at the edge and talks to the ESB.
Predicting that is one thing, but we say what we would need is to take our ESB and make it go towards a federated approach. We need to make our ESB a federated ESB such that by the time we get to that point the mini-ESBs will know how to talk to the big ESBs both when a connection exists and when a connection doesnt exist.
Thats the same interesting problem domain that Google had when they were developing their technology for applications. Its easy to build an application that runs connected. Its hard to build an application thats sometimes connected and sometimes not and does it really well.
So are you going to get into the hardware business? Or are you going to seek partnerships?
Do I see us getting into the hardware business? Probably not. Do I see partnerships? Absolutely, yes. There are people who do this really well. We already have partnerships with some hardware vendors such as Cisco and Intel.
Well, Microsoft has gotten into the hardware business with things like the Zune and Xbox.
Well, yes, and you can take lessons from what they are doing and also what Apple has done with the iPod and iTunes, etc.
In that regard that strategy is similar to ours, in that they have a platform. We want to be the platform for the enterprise. Where do we put the platform? Its not about the cool technology, which we do. Its about taking that technology, creating a platform that other people support and then connecting that technology to as many things as an ecosystem to make it happen.
What is BEAs response to the SAAS, or software as a service, wave?
Lets separate SAAS from a technology perspective and a business model. As a business model, I think it has proven itself. Its sort of the ASP [application service provider] Part II.
Two things happened to make it successful: One is the abundance of broadband. And two, a change in paradigm that says I can and will put my companys most intimate stuff in the hands of somebody I dont really know but I just trust. Salesforce.com has proven that that can be done.
If our customers want to provide their services as a SAAS offering, what would the platform change be? So from an innovation perspective, this is not really innovative, but a natural evolutionary step of how do I make WebLogic support building Java applications in such a way that they can be consumed in a SAAS business model?
I think SOA is helping a lot because you will be building composite apps that are easily consumable from the outside. And when we move to this world of composite apps, that lends immensely to an SAAS business model of delivering those.
Do I think that BEA will deliver its own application as a service? Yes, to some degree I can see some. Maybe in the WSIP [WebLogic Session Initiation Protocol] side, the telecommunications side, I can see that. But, in general, were a platform company. I dont think well deliver our stuff as SAAS. We will be the platform that will allow our customers to deliver their stuff as SAAS.
So youre more of an enabler?
We are more of an enabler. I would have thought that in the middle market—SMBs, etc.—companies like SAP and Oracle would actually find it useful to deliver their software as a service. Because one of the inhibitors to the implementation of any large-scale application—whether from Oracle, Microsoft or SAP—is the complexity and the commitment you need to give it to start. Theres a whole slew of [companies] that just cant afford the complexity. That is exactly how Salesforce became successful: You dont have to worry about how I buy and build my sales or CRM application; all you do is pay us $50 a month and well worry about that.
I dont know if we and SAP and Oracle would be in the same boat. We are true enablers; they have an inherent interest in considering SAAS as a delivery mechanism for having you buy PeopleSoft access instead of buying PeopleSoft on your machine.
What does BEA have that your primary competitors dont? Why should a customer come to BEA?
The first thing is where we are in terms of getting things to market. If a customer wants to implement things now, the difference between us and some of the competitors is were there with products rather than talking about it. They work. We build only that.
And in terms of predicting where things are going and coming up with new products that enable our customers to explore new businesses, you really need to have a comprehensive, complete platform. We have that. Its about completeness.
While BEA realizes that consumer technology is entering the enterprise and has acted on that, the consumer-focused Web 2.0 companies are now realizing they may need to focus on the enterprise for growth. How do you view that?
Ill predict something from the office of the CTO. I predict that youre going to find the first wave of what we call the Internet companies—like Yahoo, etc.—in the next two years pulling high-level executives from the hard-core enterprise-level application world, like the BEAs, the IBMs, the CAs, Symantecs. And thats the sign that they realize that they need to get into the enterprise because theres money to be made there, too.
Its not just about how many people can be logged into Facebook; its can I make Facebook what everybody in the enterprise uses as part of their collaborative effort?
We recognize there is no such thing as a consumer product and an enterprise product. All consumers, when they go 9 to 5, live inside the enterprise.
How can you help expand the role of IT in the enterprise?
IT is right now the enabler of business applications. I think IT can become the enabler of business innovation. When you start giving the businesspeople tools not just to run the application, but take some of the social stuff that were bringing into the market … I think IT can use this to make people in the organization collaborate with each other. And when you start harnessing expertise across the company, youd be amazed what can happen.
I think were going to see an enormous explosion of new, successful business models in areas we dont expect because we are now building the agility into the system that will allow them to do so much more. I think of it as like giving Mozart an electronic keyboard.
Check out eWEEK.coms for the latest news, reviews and analysis in programming environments and developer tools.