By traditional management software, I assume that you are referring to software that is designed to manage the complex IT infrastructure that already underlies most services in a SOA. This type of software is commonly associated with the leading enterprise management vendors such as CA, HP, and IBM. When that sort of question is asked, there is often an implied ancillary question. It is often the case that the questioner has heard of new management products that are specific to Web services and SOA. It is natural to ask if these apparently more specific products provide any unique properties or capabilities that existing IT infrastructure management products lack or are unable to acquire.
Experience has taught us some important lessons about managing Web services and SOA, but also about operational management in general. Historically, every new advance in technology and architecture has also resulted in the introduction of new management solutions by both enterprise management vendors and startups that are specific to that new world. There is always a need for new techniques and instrumentation suitable to changes in architecture and technology. Web services and SOA are no different, and along with the technology that supports them, they represent a distinct and important new layer of IT that must be managed.
Even established enterprise management vendors who have addressed all of these earlier layers are well-aware that SOA and Web services are the new and very important layer on top of the cake. As a result, the leading management vendors have responded. For example, the Wily Technology Division of CA provides a wide range of IT infrastructure management capabilities for many layers of the existing IT stack such as J2EE and .NET based platforms. However, the capabilities of their "traditional" solution were extended some time ago to SOA and Web services through additional products that are specifically designed and instrumented for SOA and Web services such as Unicenter WSDM (Web Services Distributed Management) and the Wily Web Services Manager (WSM). The other established enterprise management software vendors such as HP and IBM are also taking this market very seriously and so "traditional" management solutions are fully capable of providing all necessary functionality in an SOA because they have or will have soon have new components as part of their complete solutions that are specifically designed for such an environment and that are designed in accord with service-oriented principles.
All that said; the above question invites an interesting corollary question. Are standalone products that only function in a SOA and that are largely intended for Web services capable of meeting all the real-world complex requirements of the enterprise SOA? There is no doubt that specialized management products can provide essential and unique value to many deployments of Web services and to SOA projects. In fact, it is fair to say that such products are usually the best first step to take. But, as small-scale initiatives evolve into enterprise-scale, mission-critical production environments, the requirements for effective management begin to change in fundamental ways. What changes is the scope and nature of your management needs. These changes don't eliminate the need for a specific SOA and Web services management product, but they do change how that solution must work with your existing enterprise management software and need to be considered when planning and deploying the earliest stages of your SOA and Web services management strategy.
Most services are themselves merely thin layers on top of already established business logic and its supporting application infrastructure – the many layers of the IT technology stack such as the various .NET and J2EE based platforms. If the layers of the IT stack below the services layer are a black box to the management software, then the management software loses sight of the business transaction. Losing sight of the business transaction or having to use multiple management tools and then trying to manually correlate the disparate information makes it hard for IT to triage problems and to meet its service level obligations.
There is no doubt that history again serves as our guide here. Again and again, the most successful management products for new technologies have evolved to become tightly woven into their associated, existing enterprise management solution. The reason is simple. Real visibility and control of your automated business processes is never just about the database table, the EJB, the Web service or any other individual entity or abstraction within the IT environment. Business and its customers don't care about those things and that is not what they hold IT accountable for. What really matters is the business transaction itself.
Ultimately, the required focus must be broader than your SOA and Web services management product. You must consider the complete business transaction from its initiation at the service consumer stage down to the deepest levels of your IT infrastructure. Your SOA and Web services management product must be able to interoperate in a deep and meaningful way other elements of your complete enterprise management solution. In fact, one could say that the real purpose and trend in the evolution of management software is not that we monitor different things than we used to, or that we may use different protocols and standards to support the operation of the management system over time, but that we have become aware that we must serve our customers by obtaining visibility into the business transactions as a whole that drive our business.
Dig Deeper on Topics Archive
Related Q&A from Paul Lipton
Paul Lipton suggests a couple of BPM frameworks that are available in the public domain. Continue Reading
A SearchSOA.com member asks one of our experts, "What is the difference between SOA governance and Web services management?" Read the response or ... Continue Reading