X
Home & Office

The Do's and Don'ts of Web Services: Prepare Now, but Don't Expect Returns for Awhile

The hype of Web services has subsided. However, anyone who has been around long enough to live through years of enterprise application implementations will discover that Web services and service-oriented architectures are really no better or different than the largely failed evolution of Common Object Request Broker Architecture (CORBA)- or Common Object Model (COM)-based applications.
Written by Eric Austvold, Joanie Ruf , Contributor

The hype of Web services has subsided. However, anyone who has been around long enough to live through years of enterprise application implementations will discover that Web services and service-oriented architectures are really no better or different than the largely failed evolution of Common Object Request Broker Architecture (CORBA)- or Common Object Model (COM)-based applications.

The Bottom Line: Most of the real economic benefit from Web services is years away.

What It Means: Standards for Web services are still developing, with no standards-based products available today. Return on Investment (ROI) has yet to be proven. And most of today’s projects are simple examples of the technology and lack the security, reliability, performance, and business semantics to work with multiple generations of large-scale enterprise software applications.

With CORBA and COM approaches, vendors got together and created technically elegant but impractical frameworks for implementation of heterogeneous applications. Each required too much homogeneity for practical real-world use beyond a single application footprint or the occasional private, single-company secured network deployment of a small-scale set of applications.

While Web services are a step forward, technology buyers are still left with unanswered issues related to broad adoption of Web services. Questions abound about the following:

  • Security
  • Guaranteed reliable messaging
  • Change management
  • Performance
  • Semantics
  • Service-oriented architecture
No CIO has a clean whiteboard of technology to start with, since everyone must deal with two to four generations of old technology that will either have to be rewritten or substantially upgraded to fully take advantage of Web services.

Recommendations: So what should clients do when it comes to Web services? AMR Research offers the Do’s and Don’ts of Web services for technology buyers.

Do:

  • Participate in standards committees (W3C, OASIS, WS-I) by offering real-world scenarios that can be used as test cases for ensuring reliable, secure, and high-performance Web services.
  • Participate in semantic committees (such as RosettaNet, ebXML, cXML and others) that are working on the descriptions of business documents for use in Web services.
  • Test Web services technology, but don’t overpromise business results on unproven technologies that are based on moving standards.
  • Get your house in order. Continue to invest in common integration frameworks for integrating the applications behind your firewall, establishing a system of process for key business processes like quote to cash, concept to product, and others.
  • Use Web services for point-to-point integration of applications in a secure private network setting.
Don’t:
  • Wait for vendors to define the market for you; voice your issues through the standards committees.
  • Make Web services the process backbone for a service-oriented enterprise; it is not mature enough to be the backbone today.
  • Rely on Web services technology for transaction-oriented communication between enterprises until the security, reliability, manageability, and semantic standards are better defined.
  • Forget about Web services management, usually the most underestimated component of a Web services deployment.

AMR Research originally published this article on 7 October 2003.

Editorial standards