APM software improves interaction between applications and network components that support them, giving network administrators more insight and visibility into their IT infrastructures.
One of the biggest challenges facing IT departments today is delivering higher service levels for critical applications. Undiagnosed problems associated with poorly performing applications can cause unnecessary staffing and hardware upgrades, unachieved business goals and missed opportunities, all of which affect a company's bottom line. To meet these challenges, IT departments rely on application performance management software to improve services while reducing infrastructure costs and management overhead.
APM software improves interaction between applications and network components that support them. It gives network administrators more insight and visibility into their IT infrastructures. This knowledge helps them better manage application performance by identifying potential bottlenecks and fixing them before the problem areas degrade service levels.
The software, which can detect current, past and future application bottlenecks, drills into the application tiers to find a problem's root cause. It can improve application and end-user productivity by identifying and solving problems before they become severe.
Detection and diagnosis
Today's application environments are increasingly complex, multi-tiered, distributed infrastructures. APM works in a variety of application environments to provide information about supporting tiers, which can be a mix of Web, storage, application, database, network and client servers.
APM detects bottlenecks through low-overhead agents that are installed on each server and supporting tier to be monitored. The agents gather applications' performance metrics in real time and store them in an APM database to measure the baseline performance of the various functions of the application environment. The baseline information then can be used to monitor performance against desired service levels.
When application performance begins to degrade, APM software detects the problem and notifies IT staff via pager, telephone or on-screen network alert. IT staff then can use a Web browser-based GUI to navigate the alerts page, through the tiers of the application, to the lines of code where the problem typically resides.
Typical problems APM detects include database-tuning issues, poor server response time, poorly written software code and problems communicating between application tiers. Once the cause of performance degradation within a technology tier is identified, APM technology can generate tuning advice, based on the application's historical performance, to resolve the problem. For example, creating a database table index might solve the problem of time wasted on full database table scans. Or, the tuning advice might show that connection pool size between a Java 2 Platform Enterprise Edition tier and a database tier needs to be increased. The information is stored in a performance warehouse on the APM server, and serves as the basis for reports IT administrators can generate for capacity planning.
Because performance problems tend to come and go as a result of the nature and volume of transactions, IT staff needs to be able to analyze past performance as well as real-time performance to understand trends. By doing so, IT departments can take a more preventive approach to managing performance bottlenecks, which can drastically improve their companies' overall investment.
With business applications seeming to grow by another tier every year, APM will soon go from IT requirement to data center necessity. Future uses of this technology will extend its visibility into application performance and usage patterns, driving business efficiency through enhanced management of asset utilization, application usage and cost allocation.
Murphy is director of application performance management product marketing for Veritas Software. He can be reached at firstname.lastname@example.org.