Most established organisations have some legacy mainframe systems that are still running business-critical applications. Perhaps it is a billing system for a utility company, a bank processing application, or a weight and balance application for an airline.

In many cases, these antiquated relics have outlived the companies that created them and the PL/I, COBOL, Pascal, or FORTRAN programmers that wrote their software. Yet companies still rely on these systems because developing replacements with modern software architectures can require years of effort and millions of dollars.

Most of the time, these systems continue to work fairly well. But when problems occur, the IT administrators responsible for the performance of these systems have few tools at their disposal to troubleshoot and fix the problem. Application performance management (APM) for legacy mainframe systems does not exist today outside of rudimentary tools that monitor the availability of specific components or homegrown tools that are elegant solutions for a very narrow set of tasks. This is not a criticism, but simply a fact of life.

Organisations that rely on these legacy mainframe systems can suffer keenly because of this lack of visibility into performance. Fortunately, there is now a solution to this acute problem. New network-based application performance management (APM) technology can provide insight into the performance of these applications as long as they communicate across a network. The reliable nature of the transmission control protocol (TCP) enables network-based APM devices to extract a wealth of health and performance metrics for legacy mainframe applications.

An example of how network-based APM can provide critical visibility is from a utility company that was running an old Stratus VOS application for its wastewater operations. The application had experienced a slow degradation in performance that threatened the company’s regulatory compliance. As its legacy system slowly ground to a halt, the company risked penalties from regulatory agencies because it could not report quality control readings in a timely manner.

The utility company deployed a network-based APM solution and quickly detected and resolved a number of performance issues, including misconfigurations that caused the system to send a number of extraneous DNS requests that subsequently failed. Because the IT team could see all client-server communications, including DNS errors, they could easily fix the problem with a simple host entry.

Not only did this fix increase the scalability and longevity of the legacy system, but it also dramatically alleviated the burden on that company’s DNS infrastructure. In fact, the company had planned to spend over U.S.$150,000 on new servers to upgrade its DNS infrastructure. With the DNS issues related to this legacy system now resolved, the additional servers were no longer necessary.

In another case, an airline used a network-based APM solution to troubleshoot performance problems with its weight and balance application. After each airplane is boarded and luggage stowed, the weight and balance of the aircraft must be calculated before the plane can taxi to the runway. Slowdowns in performance can cause delays—and angry passengers!

The new monitoring and troubleshooting solution helped this airline pinpoint and remove bottlenecks in their weight and balance application and contributed to that company’s move from seventh to second ranking in on-time takeoffs among its industry peers.

In addition to gathering real-time performance metrics for DNS and TCP, network-based APM solutions can gather metrics for application transactions that use other L7 protocols as well, including HTTP, database, and storage protocols. These solutions are deployed as either a physical or virtual appliance connected to a network tap or port mirror, and analyze application transactions passively, using a copy of network traffic.

This non-invasive deployment requires no configuration and does not impose any performance overhead on system resources, as is the case with host-based instrumentation used by most other performance monitoring tools.

The point here is that organisations that still rely on unsupported legacy systems are not entirely bereft of performance monitoring capabilities. New APM products that use network traffic can restore real-time visibility into the performance of these systems and even provide trend-based alerts for anomalous behavior.

This is crucial for companies who still run business-critical applications on these systems but currently do not have a more modern replacement available. While the ideal solution would be to replace the legacy technology, a network-based APM product can be deployed easily in the meantime to provide critical real-time performance monitoring and troubleshooting capabilities.