WAN Delay Analysis

The subject company’s current use of the Event Processor and Troubleshooting toolsets is best described as reactive management - when services fail and when end users report poor performance (slow response time), the staff tries to determine the cause. Reactive management is a necessary evil, but can be complemented by other activities to run an enterprise network more cost-effectively. During our analysis, we illustrated that certain Visual UpTime™ Troubleshooting tools are best applied pro-actively.

While user perception of "response time" is a driving factor, this perception is highly subjective. Many enterprises develop custom tools to measure application delay. During our interviews, we learned that the subject company is reluctant to invest in custom software, but wanted to get a better sense of network response times. We also note that centralized custom tools introduce traffic onto the WAN, and because they are run on protocol analyzers, PC’s or workstations, they actually measure delay between hosts rather than delay between Frame Relay access lines.

The Visual UpTime™ Troubleshooting toolset provides the subject company with a Round Trip Time (RTT) Measurement tool that does not require any investment in custom software. WAN delay analysis can be performed by correlating frame transmission and reception from ASE’s on both ends of a PVC, resulting in a protocol independent, non-intrusive analysis which measures only that portion of the application delay that is attributable to Frame Relay.

We believe the appropriate application for the RTT measurement feature of is a two-fold practice of:

1. asserting a baseline by measuring current response times which appear satisfactory to customers, and
2. (2) periodically monitoring Frame Relay to confirm that this baseline continues to be met.

This information will prove useful if the Frame Relay service degrades over time, and will be extremely valuable when the subject company finds it necessary to reevaluate and make changes to network connectivity. We also believe that one of the more promising uses of this tool for the subject company is to determine whether the propagation (emission) characteristics of DDS access lines for broker locations (the bounding transmission channel in this network) are sufficient to support applications with more stringent latency requirements than the current (broadcast delivery) application.

next...