February 06, 2001, 3:53 PM — Traditional management tools that rely on post-event warnings just don't suffice in the real-time e-business world. By the time you're aware of a problem, your e-business customers are already gone.
What's needed is an early warning system that alerts you to potential service problems, even those caused by non-e-business processes running within your shared IT infrastructure.
In today's e-business infrastructure, problems can occur at any point across a transaction's critical path, in a single component or between multiple components. Yet traditional management tools are not designed to monitor this cross-component transaction flow, nor are they designed for real-time early warning.
But a new class of technology called real-time management makes it possible to monitor the flow of e-business transactions or information across networks, servers and applications in real time.
Real-time management systems (RTMS) require three elements: data collection, analysis and reporting.
- Data collection. These systems gather resource-utilization metrics of heterogeneous components across e-business infrastructures. Only after the data has been collected and analyzed can meaningful decisions be made about strategies for optimizing performance.
For example, with an online trading site you would have to collect data from many sources, including real-time stock quotes and customer portfolios stored in databases. To monitor service levels in real time, critical components -- including network devices, Web application servers and transaction systems -- must be monitored and input must be collected. Agents on each of these components gather critical information in real time and report it for simultaneous analysis.
An RTMS also interfaces with legacy management tools. Data collected from other vendor's products such as IBM Corp. Remote Monitoring, which measures utilization of mainframe resources, or NetScout Systems Inc.'s Ethernet Probes and Token Ring Probes, which provide network performance information, can easily integrate with the collected data of an RTMS.
An RTMS continuously monitors service and consumption metrics and sends details to a correlation engine. To do this effectively it must be able to aggregate large amounts of heterogeneous data, and immediately analyze this heterogeneous base for trends, relationships or early warnings of service degradation.