eWEEK Labs recently conducted a roundtable discussion on Web services. Moderated by Technology Editor Peter Coffee, the roundtable comprised key players in several areas of the Web services space. The discussion centered on where Web services are as a technology, where the industry is in achieving effective consensus on standards, and how the business models are developing for using those standards and technologies.
Weve heard a lot from Microsoft during the last year about the rebuilding of its development suite and the whole Microsoft platform around more of a services model with the .Net Framework. Where are you in that progress?
Charney: Microsoft has been thinking and working around Web services for quite some time now. When we announced our .Net initiative, one of the biggest challenges was that it was the first time we, or many people in the industry, were even talking about this concept of Web services. Since that time, a few years ago, weve been implementing Web service support across everything weve done and everything were doing moving forward.
Where is Oracle on the migration toward XML as almost the native format of data?
Farrell: Oracle believes in the database for storage of things, and weve added the XML capabilities to the database to allow users to actually store XML components and objects in the database and access those with the same systems that they are familiar with.
As far as enabling everything else, the real object is to be able to get the data and enhance what you have and expose that as Web services. So, rather than taking the approach of rewriting things in Web services or rewriting things based on Web services, we focus on taking applications that our customers have and the infrastructure that we provide and putting in the hooks for people to be able to expose different functionalities and data as Web services.
Are you finding it necessary to make choices between different approaches, such as the Java platform versus the .Net platform?
Farrell: We do have to give consideration to the different run-times, but, in general, one of Oracles strengths is the integration and the different flavors of data and information that we actually work with. So, whether its a .Net Web service or a .Net architecture that theyre building on or Oracle J2EE [Java 2 Platform, Enterprise Edition], the database is there to support it. Obviously, were not a .Net application server but a J2EE application server, but we can consume and talk with and orchestrate with .Net Web services, and all the data can be stored in an Oracle database.
IBM has a really heterogeneous lineup of server platforms, in terms of different technologies at the core of the server aimed at different tiers of the enterprise. Are you finding Web services an effective way to present that as more of a unified offering to your customers?
Norsworthy: Certainly, IBM has a long history of delivering heterogeneous solutions, and we do that not because we decided to but because its what our customers want. So, Web services is a great enabler for our customers, who, as you point out, have a lot of diverse systems, both hardware and software, to enable them to speed their ability to integrate across those systems.
The BEA platform has been a key target for application developers. What leverage does an application server layer like BEAs and the associated tools provide?
Renaud: The notion of an application server has been evolving very rapidly toward the application platform suite. The application platform suite basically acknowledges one of the basic realities of development—that developing an application without leveraging existing software assets, without doing some application integration in the process of development—is not nearly as interesting as being able to develop your entire application to take advantage of all the stuff that you already have running.
WHAT MAKES A WEB SERVICES STRATEGY?
Where does Sonic fit in to the Web services ecosystem?
Van Huizen: Sonic ties together applications that have been exposed [as] services. So what weve been spearheading is a new kind of middleware offering called the Enterprise Service Bus, which takes into account that, over time, applications will be built to integrate, will be built to interoperate. They will have the service interfaces available, and theres a new kind of integration infrastructure that is required to tie them together in a broad way. So we focus on the interaction between applications, how that is hosted, how its managed, how its secured, how you make it reliable and how youd go about creating a global infrastructure of collaborating applications.
Where does Systinet fit into all of this?
Bjork: Systinets focus from the beginning has really been on providing cross-platform interoperability for developers, in terms of Web service enablement for their platforms. One of the things that were seeing quite clearly in the market at this point in time is that SOAP [Simple Object Access Protocol] enablement is really one of the first key industry drivers that all vendors have been able to agree upon, in terms of putting it into their products. What were seeing now is a proliferation of that throughout application developer products, and I think, longer-term, Systinets strategy has really been focused on providing the tools and technologies that allow ISVs and OEMs to embed that kind of service technology within their platforms. More and more, I think youre going to see enterprises getting introduced to Web services technology through application providers like Interwoven [Inc.] and Documentum [Inc.] and others.
How do you see Avanades role as a non-Microsoft developer relying primarily on the .Net technology?
Hartman: As were positioned to do Microsoft-based system integration solutions, were often called upon in enterprise scenarios to build solutions that interoperate with other technologies. Weve got quite a few customers in production already that are basing Web service solutions on top of mainframes and interoperating with J2EE, so I think a lot of the promise of Web services are truly coming to bear.
So really theres no such thing as being a Web services technology company focused on .Net or J2EE, for example, because by being a Web services developer, you accept the need to integrate all of those legacy assets and to take advantage of all of the new technologies?
Hartman: Absolutely. Its not just about a particular platform choice of J2EE versus .Net, versus the mainframe, versus anything else; its about how youre going to pull these pieces together through integration and interoperability to solve a real business problem.
How do you folks feel about the nature of the Java process at this point and whether its moving forward in a way that makes many of the main IT players important partners in that development?
Farrell: I think that there has been some talk about, and ongoing concern about, the ownership of Java. I think, when it comes down to it, theres also a lot of talk around standards and what that word actually means. One important thing to point out is, “standard” to us is how many people actually are using and adopt something. I think one thing that J2EE and Java have done very well is gain the adoption of a very large group of customers and vendors in the marketplace.
WHATS THE WRONG WAY TO THINK ABOUT WEB SERVICES?
I think it works very well in its goal, which is having all of us—for the first time in history, really— all of the major companies, except for a few, on the same platform and working together to better that platform.
IBM has made a lot of investments in Java. How do you feel about IBMs role in the process of improving and refining its definition?
Norsworthy: Weve invested in Java from very early stages, and I think the point is right on that whats important here is the level of customer adoption we have, which has just been phenomenal. So, of course, the road going forward will always have a couple of turns in it, and well continue to work to evolve that process to be better and better. But IBM and its customers have a significant investment in Java thats been successful for all of us.
What made .Net necessary, as opposed to allowing J2EE to be even more pervasive than it is?
Charney: To go back to the point that has been made throughout the conversation so far, one of the key elements of Web services is this notion of heterogeneity and connecting various systems together. Thats one of the key areas that our customers have made very clear—that theyre going to make investments in a variety of systems and a variety of platforms. So whats important is that the vendors are aligned around a way for those different systems and applications to communicate with one another.
One of the key elements of .Net is support for standards—such as WSDL [Web Services Description Language] SOAP [Simple Object Access Protocol] and XML—as key models for interoperability. Certainly, Java has a challenge in the sense that it has not yet standardized around a Web services model. I think its going to get there, but—and I think rightly so—none of the vendors on this call is waiting for that to happen.
Whats the distinction between an application built around a Web services model and an application that can send and receive SOAP messages and exchange XML-formatted data?
Charney: When youve architected an application around Web services, its really built for that. So what that translates into is significantly less lines of code in order to get something exposed as a Web service. The challenge is when something like J2EE really wasnt designed and built around a Web services model—its sort of an after-the-fact integration.
But, at the end of the day, whatever the customer decides on implementing, the thing we cannot disagree on—and we should all be consistent upon—is that the language of Web services is consistent. And I think thats where organizations like WS-I [Web Services-Interoperability] and organizations that are committed to helping make sure that that interoperability exists come in.
Norsworthy: One of the key messages here that we must continue to emphasize is that Web services and J2EE and .Net are not the same things. J2EE and .Net are the same, but Web services is something entirely different. Web services ensures interoperability, but it does not affect platform choice because what people look for in a platform is investment protection.
Weve heard from a number of prospective enterprise users that Web services technologies are seen as a cost-effective means of application integration and of getting synergy across different platforms and different families of technology. However, these prospective users are also quite skeptical that the technologies are as secure and robust as they need to be.
Farrell: We have customers today using Web services in real-world situations—very secure, highly available Web services. I think what were missing today is the standard approach to that. Weve obviously implemented something because the standard wasnt there. WS-I and others are working toward building the Web services security standards and the other standards that you actually need for reliability on top of that, but that doesnt mean that the base core fundamentals of Web services arent ready. The concept does work.
Renaud: Like any new technology, things always start inside because some pieces are not quite done yet. Certainly, until recently, Web services security wasnt really baked. I think the base is really there, allowing you to do a lot of stuff in environments where youre not necessarily secure—for example, behind the firewall, and thats where a lot of the new stuff is started. There are some point solutions that allow you to extend out, but to get the truly fully standardized, high-level features like security, like reliability, I think were seeing a lot of progress there. In fact, we will be working with IBM and Microsoft specifically to put specifications out that will get those features out to market quickly and effectively.
I think the message to the enterprise and to the CIO is that the specifications are there, the implementations are following shortly behind and it behooves people to start thinking about how theyre going to be using [Web services].
If you got a chance to buttonhole a CEO or a CTO in the elevator, what would you want to say to them about Web services?
Charney: I would want to make sure that they understand that XML and Web services are becoming the language of business and that deciding not to invest and go in that direction would be similar to deciding not to invest in having a telephone number for your business or even a Web site for your business. Ultimately, its a belief that customers will find you, businesses will work with you—it will become an expected way of communicating and interfacing with your businesses. Thats just step one.
Looking ahead, I think its important to understand the opportunity for differentiation, and while theres a lot of debate in the industry these days that IT doesnt matter—its simply a utility, mine is the same as yours is the same as anyone elses—at least from a Microsoft perspective, we firmly believe that IT is a source of differentiation, competitive advantage, and that the companies that recognize that and that can leverage Web services as a model for providing unique services to their customers and their partners will be uniquely situated as the economy starts to improve.
Norsworthy: Id tell them this doesnt require a revolution—this can be an evolution.
Bjork: The other thing I would add is that if theyre not thinking about investing in XML Web services today, they should just go and look in their environment. Within the next six to 18 months, I can almost guarantee you that generally 100 percent of all enterprises will have a Web service-enabled platform somewhere in their enterprise.
And what will that platform do for them that their existing IT base does not?
Bjork: It will give them the capability to publish to their business partners functionality within their enterprise that they couldnt before. Web services gives them the capability to repurpose the competitive advantage of their IT assets.
Renaud: The message would be simply, think of when you develop new applications, when you support your business, think of development and integration as having the single most complexity. Think of integration—both with your existing internal assets and with your partners and your businesses out there—as something that is becoming an order of magnitude cheaper, and think about what kind of business opportunities that opens up that werent affordable before.
Hartman: [Id make] similar points to a CIO/CEO—that you have pain points today integrating with your customers, partners or even your internal systems, and you have already adopted Web services, most likely, and you may not know it. Were beyond the plumbing stage; now its time to look for new ways to drive innovation, efficiencies and opportunities that create business value for their enterprise longer term.
Van Huizen: Id like to talk to the CIO after these guys have spoken to the CEO. Or it could be a message to the CEO to reinforce with the CIO. Its definitely true that applications will be built to integrate. Its definitely true that itll be much more cost-effective and much easier to connect applications across the enterprise, and its time to begin considering what the infrastructure ramifications of that are—that there is an infrastructure that needs to be in place, that pulls all this together. It wont come for free, but itll be a lot less expensive than traditional integration approaches. Start thinking about the consequences of that and building out for it.
What are some of the basic components of that infrastructure that you feel are absent or are not sufficiently robust to do what youre talking about?
Van Huizen: A global communications infrastructure that actually connects these services together, dealing with administrative and security boundaries that are already present within the enterprise, ability to get a global view of service interactions across the enterprise and effectively manage such an infrastructure.
Given the environment as it exists today, given what you see out there, whats the mistake thats most likely to be made by an enterprise that believes it needs to adopt Web services? And what is the best way to avoid making that mistake?
Bjork: I think the biggest mistake that we could conceive of—and I think its probably one that gets made fairly consistently with new technology—is when youve got a hammer, it all looks like a nail. Bite the stuff off in chunks, get definable projects where you believe there could be a rapid return on investment in terms of what youre trying to get accomplished with Web services, implement a project and then measure the results of that project. If youve gained what you should have been gaining—and we certainly believe you will—then move it on to a broader rollout and certainly look at an enterprise-level [rollout] at some point along the way.
Farrell: One thing we didnt talk about today or that was just hinted on briefly is management of all this. Now I have these Web services and theyre running. How am I keeping track of them? How am I managing that?
At Oracle, for example … were extending functionality and adding Web services. Our Enterprise Manager, for example: Its the same program you use to manage your database clusters, the same program you use to manage your application servers, and your Web services are right there, too.
Charney: I think one of the mistakes would be to assume that Web services are not there yet and its OK to sit this one out and wait and see what happens.
It is time to start understanding that model and the level at which you invest in it—the depth and the breadth is really up to you and what makes sense for your business.
Norsworthy: I think we all sort of agree. You start small, you grow fast and you use WebSphere [laughter from all].