Top Links You Must Click On
Java Industry News
Exclusive 2004 Interview with Thomas Kurian | @GoogleCloud #CloudNative #DevOps #Serverless #Kubernetes
Databases, Middleware, The Grid, and More...
By: Jeremy Geelan
Jan. 28, 2019 06:15 AM
March 5, 2004
Two and a half years ago, when he stepped in to help Oracle get back into the application server business, Thomas Kurian had quite a tough row to hoe. Now he has 19,000 customers. JDJ explores with Kurian his 30-month journey, and asks too about other aspects of emerging technology seen from the Oracle perspective.
In JDJ last month, Sun's Jonathan Schwartz claimed the "death" of middleware, so my first question to Thomas Kurian is inevitable, given his role as a champion of middleware for Oracle. What does he make of Schwartz's claim?
"Those who say such things are being facetious," replies Kurian, bluntly. "A company that says that there is no role for middleware means it in the sense that it feels its own operating systems incorporate middleware facilities. Such a company is bundling proprietary middleware into the OS as a vehicle to try and lock customers in."
Kurian compares the current situation vis-a-vis middleware with relational databases in the early days.
"Early on, relational databases were primarily just query engines, and the information management infrastructure was very rudimentary," he explains. "Over a 20-year period, relational databases have incorporated a number of specialized features and functionalities - clustering capabilities, data warehousing capabilities, security for the data, and so on - and today you might argue there is certainly a lot of change still in database computing, but people see it as a much more mature infrastructure for information management.
"Now with first-generation middleware," Kurian continues, "the killer application was to deploy processing logic into a J2EE server and scale up the application so lots of clients can access it."
Now he believes that what we're seeing is an "evolution" of middleware.
"During the early phase of such an evolution obviously the rate of change is higher, but over a period of time these technologies mature and at some time in the future - in two, three, five years, I don't want to make a prediction - there will be a much more mature application infrastructure, and while there will continue to be change it will be at a lesser rate than today."
Does Oracle see greater opportunities in this mature phase, I wondered.
"Part of the process of maturation," replies Kurian, "is that customers want to eliminate all the piecemeal componentry that they had been using and shift to a more mature platform that offers a broader set of services.
"When we entered the market two and a half years ago with our integrated application server we were the first to integrate all these pieces into one app server. That to us is the biggest opportunity from a business point of view. The customers that we talk to validate this again and again - telling us how they save cost and time, either from a development perspective because they can use a single development tool and framework to do a lot more things, or from a deployment and management perspective because they have a much simpler and easier-to-manage infrastructure. It saves them a lot of cost and gives them a lot of value and that's why we continue to invest so heavily in making this vision a reality."
Can the Oracle vision be unpacked in ways that differentiate it from the other giants, like IBM, BEA, and Sun? Kurian was very precise when talking about the various distinguishers.
"When you look at Oracle Application Server versus IBM," he says, "our middleware has always been built as a single cohesive development tool with a single cohesive runtime platform, not a brand like WebSphere with many different pieces, but a single integrated product. It is this degree of cohesion that is the big difference compared to IBM.
"Compared to BEA," he goes on, "the biggest difference is the breadth of our capability. We are broader in two dimensions. First, in terms of deployment and management: we have much more sophisticated capabilities in the way we can manage the entire application server and all of its services - monitoring, life-cycle deployment, software provisioning, identity and security management - we make all of that much easier. Second, we have more capabilities than WebLogic around Web services, wireless access, business intelligence, and so on."
And what of the last of the Big Three rivals, Sun?
"Compared to Sun, the first big difference is that our application server is actually used by lots of customers in a heterogenous environment, not just on Solaris but on lots of different OSes - Solaris, HP, Windows, name your favorite. They access Oracle and non-Oracle databases; they use it to access mainframe systems; they use it to access applications like SAP. We have 19,000 customers using it in a highly heterogenous, multivendor environment while Sun typically pitches its product at Solaris or Sun shops."
In the pecking order, Kurian says, Oracle feels it's in second place with the number of customers - with 19,000 plus, whereas BEA has only about 15,000.
"In terms of number of new customers," he adds, "every quarter we add 1,500-1,700 new customers, pretty consistently. So we feel very comfortable that in terms of real deployment we're number two in the market."
There's some growth in the overall market, he says; application servers didn't decline, but they were pretty flat during the recent tough market conditions. The downturn had an impact on smaller companies, naturally.
"The paradox is that as the new versions of J2EE standards come out - 1.3, 1.4, 1.5 - J2EE is encompassing more and more capability to allow the Java programmer to do more things. Things that, for example, in 1.5, EJB 2.1, the JMS enhancements, give the developer more capabilities to do enterprise application integration a lot better. Enterprise portals with the new JSR standards are allowing the Java developer to solve more business problems than their companies actually have!"
Is the paradox that even as J2EE incorporates a number of these capabilities, it needs to be kept accessible to those people who want to do simple things simply?
"Yes," says Kurian, "there's a risk that J2EE gets so complicated and so cumbersome that people who want to build very simple applications find it challenging.
"That's where I believe personally that some of the new efforts over, say, JavaServer Faces help," he adds. "Compared to .NET, which spends a lot of time talking about how to build presentation-centric applications, with WinForms and .NET forms and things like that, J2EE - other than JSP - has really not had a great model for presentation-oriented kinds of applications. These are what a lot of people want to get going with; they want to build simple, forms-style applications."
Oracle participates very actively in the J2EE process, says Kurian. He has about 150 people in his organization who participated in about 139 JSRs in the last year alone.
"We feel that the Java Community Process is the right model for Java. The community has to find a way to leverage the best efforts of people and make this problem of complexity self-correcting."
Of course, the JCP is not without its drawbacks and limitations.
"Some vendors play games with the community process," Kurian concedes, "trying to take control of it and take it in specific directions in an attempt to favor their product, as opposed to making sure that Java retains both the nascent appeal that it has for people as the only solution that allows them to write portable applications while continuing to give them the ability to solve more and more relevant problems for their own organizations.
"Some of the vendors," he says, "try to kill the golden goose, altering the very thing that made their products successful - open standards - by trying to take control of these standards in a proprietary way."
Web Services - No Cure-All
"Larry has always said that Web services have an important role to play within IT organizations," says Kurian, when I ask about Ellison's perspective. "His skepticism has always been related to when vendors claim that Web services are a panacea for all kinds of IT problems, and that there's simply a magic whereby if you wrap everything in Web services, they know how to communicate with everything automatically.
"That's where his skepticism came from," Kurian explains, "and it's been justified.
"Our view of Web services at Oracle is that they have a very good value proposition for organizations, particularly in the context of taking and modernizing your existing legacy infrastructures by wrapping them with a more standardized set of interfaces so they can be accessed in a more rational way, a more standardized way, from new systems, new applications you're building, or even tied into corporate-wide business processes, for business process automation."
But there have been limitations to the way that Web services has played out, reckons Kurian.
"Where things have fallen down is twofold," he observes. "First, Web services was overhyped by Microsoft and IBM as a technology that you could use to do anything you needed, and there's always some disillusionment after such hype.
"Second, there's been a real slow process in terms of the implementation and acceptance of Web services standards, particularly in areas like reliable messaging, security, coordination and activation, process management, orchestration, as well as things like policy."
In Kurian's view, these are the two main elements that have hindered their adoption today within enterprises.
"When Oracle talks about the grid," Kurian answers, "we're not talking about a new programming model or a new computing paradigm.
"What we're talking about is how do you take your existing applications, as well as new modern Internet applications that you may have built, and deploy them at a much lower hardware cost - CPUs, storage, servers - and get the same reliability, low cost management, etc., that you enjoyed when you deployed them on big servers.
"We're not talking about grid computing as a paradigm shift in the way people ought to design their applications. The critical thing, we tell customers, is that any application that you write to a modern architecture like J2EE can be automatically deployed on the grid.
"We are addressing the central question that customers have today," he adds, "which is that while they have enormous budgetary pressure on their IT infrastructure, they have a lot of excess capacity sitting around in their IT environment."
Which makes Oracle 10g, presumably, the right app server at the right time?
Kurian smiles. "OAS 10g is the result of over 5 million hours of engineering time, and we spent a lot of time looking at the primary problems in the way they built, deployed, and managed applications. So, yes."
But his high opinion of 10g isn't formed in a vacuum.
"Over the last two years we've had more than 4 million people download our application server off our Oracle Technology Network," Kurian notes. "We delivered 10g just before Christmas, in the middle of December, and over the Christmas holidays we had close to 12,000 downloads and the feedback has been great."
Turning to 2004, I ask Kurian what kind of year he's expecting, given the givens - election year, a tough economy.
"We are definitely seeing stabilization in the market as well as an uptick in spending," he replies. "The recession has just been a lull in the process of customers transforming themselves from traditional models of business to the Internet-oriented model. That process will continue - people don't argue with the fundamental value of the new computing model.
"In our view," Kurian continues, "the reason we're really excited about 2004 is that in the next two or three years every company in the world is going to go to an Internet model of computing and every company will want to deploy applications on a server, and have the server - the application server and the middleware - provide the capability to do fundamentally new things better, faster, and easier, and to run this with greater reliability and manageability.
"We have been working at this for two years now and we think that this is only going to accelerate now that we are coming out of this tight budgetary environment. What's happened during the recession is that members of our customer community have become much more aware of how they are spending money on information technology and are much more cost-conscious even as they have more dollars to spend."
Pursuing the Dream
"It's not just my own passion, it's shared by the great team of people I work with," he points out modestly.
"I think we have one common belief, and that belief is that this shift of the computing model happens only once every generation," Kurian observes, with more than a hint of awe.
"No one has talked so far about a ‘post-Internet' computing model," he continues. In other words, the current cycle is far from over yet.
"Having the chance to help customers and organizations successfully go through that fundamental change is something that very few people get to do, and very few companies get to do. We felt that we wanted to build a software product, an application server, that was so good, so capable, so easy to use and so powerful, that every customer, every company, every organization in the world would try our stuff and like it.
"That's still our dream," Kurian says, "that's still what we work for, and that's still what everyone here at Oracle is focused on. It's just so apparent to us why we feel so passionately about enabling everybody in the world to try our stuff."
Thomas Kurian, in other words, feels almost a sense of personal privilege at being the man he is, in the job he is, at the moment he is.
"I do feel privileged, yes," he replies. "We are all privileged. We are all very lucky at Oracle to have the resources to develop a product that we can make available to so many people. We are privileged that we have 19,000 customers - soon many more hopefully - who use our product.
"It is the greatest privilege in the world to build something that the whole world is going to try, and that continues to be our goal."
With passion like this on tap, it is certain to be an interesting year for Oracle, with or without a PeopleSoft deal.
Reader Feedback: Page 1 of 1
Enterprise Open Source Magazine Latest Stories . . .
Subscribe to the World's Most Powerful Newsletters
Subscribe to Our Rss Feeds & Get Your SYS-CON News Live!
SYS-CON Featured Whitepapers
Most Read This Week