Organizations are adopting service-oriented architecture to ease the integration of applications, map business processes to software IT structures and establish interoperability among diverse systems.
Yet for all its promise, SOA also can complicate the IT environment and disrupt operations it is intended to improve.
Tools for application performance management (APM) can help IT departments master the complexity of SOA environments. Through active performance monitoring and expert tuning capabilities, APM can help them anticipate, identify and resolve problems quickly and efficiently.
A number of factors contribute to the complexity of managing SOA applications:
-- The applications typically are multitier, and all of the components and their interactions must be monitored.
-- SOA-based applications often are heterogeneous and cross physical-server or public-private network boundaries. Consequently, they must be monitored across platforms, and the extra hops make failures and performance bottlenecks more likely.
-- Because SOA-based applications are relatively new, they often introduce unfamiliar elements to IT administrators, and their use in rapidly changing environments means operating manuals, training and support infrastructure must be kept up-to-date.
-- Because SOA-based applications abstract underlying implementations, traditional performance testing during development may be insufficient to ensure performance in production. In addition, while the loose coupling of software services in SOA environments adds flexibility, it may interfere with visibility and control.
IT organizations must extend their performance-management practices to meet these more complex, demanding SOA environments where performance problems are experienced more quickly and broadly than they are with traditional applications.
To work effectively in SOA environments, APM tools should take a user perspective toward performance, tracking transactions across tiers and correlating results. They should cover Web servers, applications, application logic, application servers, middleware, network flow, database activity and all storage layers. By simulating real SOA transactions to pinpoint problems before they affect users or partner systems, these tools can help ensure that services remain available continuously.
Because SOA environments change so rapidly, APM tools should be able to analyze performance data and deliver intelligence about components and interactions. This may include, for example, forecasts about application scalability, data regarding past and current compliance with service-level agreements (SLA) or actionable advice to help remediate performance bottlenecks in SOA-based applications.
APM tools also should provide a single view of all relevant performance data while letting IT organizations build customized views into SOA-based application performance without switching between multiple tools and user interfaces.
By helping SOA applications run optimally, APM tools promote the primary goal of SOA: improved business agility. APM also helps IT departments achieve higher returns on enterprise SOA investments by making staff, infrastructure and applications more effective. Moreover, APM lets IT departments meet critical SLA requirements and resolve issues before they become problems.
The advantages of leveraging APM in SOA environments are compelling, delivering improved business visibility and control for significant gains in agility, flexibility and productivity.
Mark Lohmeyer is vice president, product management, at Symantec and is responsible for defining the product strategy for the Server Foundation and Application Performance Management product groups. You can reach him at firstname.lastname@example.org.