Creating those scripts requires extensive knowledge of scripting, formatting, SNMP and statistical analysis. And on top of that, you must analyze the data to determine whether you have enough ports to service your users. Your time is better spent managing your remote-access servers than writing and debugging scripts.

In response to this need, Concord Communications has provided remote-access tracking functionality in Network Health-Remote Access, a product aimed at easing data collection, information processing and reporting for remote-access servers. I tested a beta version of Network Health-Remote Access in Network Computing's Syracuse University Real-World Labs® and was quite impressed by its breadth of reporting and automated data analysis.

Detailed Network X-Rays During testing, I pointed Network Health at the lab's 3Com Corp./U.S. Robotics Total Control network hubs and let it gather data for five days. Network Health polled the Total Control hubs every 2.5 minutes, gathering modem and RAS (remote-access server) information via SNMP.

Like other baselining tools, such as Kaspia Systems' Enterprise Systems Monitor, Network Health must run over time to generate a baseline. Nevertheless, I was immediately pleased with its At-a-Glance reports, which offered meaningful data after just a few minutes. The longer Network Health runs, the more meaningful the graphs.

The At-a-Glance reports provide an overview of a remote-access-server element using thumbnail graphs; clicking on each graph gives you a detailed view. The Modem States line graphs show the percentage of time that the modems are in various states, such as on-hook, off-hook, busy or test mode. This information is central to identifying the busiest hour in a particular day. For detailed information on our entire modem pool, I looked at trend-oriented graphs, such as Health, Service Level and Trend Analysis.

Many of Network Health's more useful reports make comparisons between current data and a baseline, which is extracted from the data at the same time and day of the week for the same devices. After running for several days, Network Health generated more substantial trend reports. During testing, I scheduled several reports to run every morning during low periods to reduce the SNMP traffic loads on the Total Control network hubs.

The Health report presents a number of graphs about the remote-access servers' behavior. For my purposes, the Health report graphed the total call volume in total minutes used by the remote-access servers for the past seven days, while another graph displayed the call volume in minutes for the day the report ran. These two reports provided important trending information for determining when the demand for dial-up lines is highest on a daily basis.