Managing IT has always been about ensuring the availability and performance of the application. At one time, applications were simply transactions on a single system. They might have required databases, business logic and networks, but they were all on a single platform. As IT evolved, business needs drove a requirement for application flexibility, but that also increased complexity. Today, applications are often Web-enabled and span platforms, geographies, and even organizations and companies. But as the requirements have evolved, management of mainframe systems and applications hasn’t changed to match business needs.
Most performance monitoring tools have traditionally focused on a single platform or single subsystem, such as a database or transaction processing environment. These tools were never intended for monitoring complex, Web-enabled, multi-tier applications—frequently called “composite applications.” The addition of virtualized environments and cloud computing has added further complexity to the mix.
Composite applications span IT silos, so management solutions to monitor these applications also need to span silos. The current silo technique typically shows subsystem-specific status of a red or yellow light, indicating a problem. But it doesn’t tell you which applications are impacted. Furthermore, few organizations can define all the paths their transactions may take across servers, databases, and the rest of their infrastructure. Frequently, application problems result in the “war-room” scenario, where department heads come prepared with performance data that indicates the problem can’t be sourced to their area. Every minute spent trying to determine the source of an application problem tied to the mainframe or elsewhere may result in upset users and customers and lost opportunity or revenue.
At a minimum, this is an inefficient process. Even if it’s an internal application, the problem may result in prolonged productivity lapses costing the business or organization much-needed dollars. Also, think about the last few war-room meetings your organization held and consider how much an hour or two of each attendee’s time cost. Add to that the time consumed by follow-up meetings, calls, emails, etc. to close out the “issue,” and then consider what the costs and time away from more productive work add up to for your organization. Multiply that by the number of “one-off” meetings you actually need to have each year and the true cost of a silo-based view begins to come into focus.
Systems managers need a holistic view of their applications; they need to know what components each application uses. With this view and knowledge of the importance of the application to the business, triage can be performed, ensuring that business needs and costs are considered. It also means the problems can be resolved faster, which is critical, especially in the world of fickle Web users. These users—whether they’re customers or internal and external stakeholders—demand immediate, continuous access. Non-availability isn’t acceptable and even a slow system will push them elsewhere to get what they want (see Figure 1).
Ultimately, systems programmers and their management need to be able to proactively avoid problems, and when required, quickly resolve them. Arming programmers with the right tools will evolve their method of management to include a powerful awareness of the entire enterprise. Accomplishing this requires solutions that can trace an application path and then drill down into the root cause of a multi-tier application problem. Companies are addressing this need with integrated and semi-integrated solutions. You’ll need to research these and other industry offerings to see if they’re truly the end-to-end, integrated solution you need for your environment; one or more may be.
Mainframe and distributed workloads and environments are growing at a healthy pace and will continue to coexist for the foreseeable future. Business transactions also will continue to span these platforms. The new mode of operations needs to be a cross-silo, platform-agnostic approach where we care as much about the life of a business transaction as we do about each subsystem. To make this approach work, we’ll need tools that improve communications between platform experts and application support teams and help diminish the need for war-room meetings that plague so many organizations today.