Are you listening to SNMP traps? Most of the problem detection you're looking for can be done with that... fast-busies and ring-no-answers usually show up as an SNMP trap also. If for example you're running the old DSP code with the infamous modems-failing-in-pairs problem, when a customer hits a bad modem, the DSP will send a trap. (I think -- it's been a while.) Ring-no-answer definitely sends a trap, something like "inbound call failure" or "watchdog timeout" or some kinda packet bus failure. You can get traps when T1's go down or just get noisy too... just gotta turn 'em on. Of course the real challenge is finding SNMP trap listening software that doesn't suck and that doesn't cost a fortune, which at least for Win32 I have yet to find... Regular busies should only happen if you're full, and MRTG oughta be able to give you at least a rough count of used modems... for testing IP connectivity, there's always ping :) Calling out to all your dialups should be easy enough, except that your client modem isn't likely to be able to tell the difference between a fast busy and a regular busy, so you don't know if you're full or if you've got a dud DSP or PRI without checking a second source of info. It's an interesting idea though. The only problem with doing that kinda thing at the client end is that there's not a lot it can tell you when you're connected. You could maybe ping the other end and maybe get the (inaccurate as hell) initial connect speed from Windows, but that's about it. I'm not a Win32 programmer (I'm a Perl+Unix guy) so I'm not even sure how you'd get that much. There's stats you can get from the modem after you hang up, but to get that your program would have to know about every common chipset out there (USR, Conexant, Lucent, Motorola, PCTel, Ambient, Multitech...) and the AT commands to pull the stats up. Mike Andrews (MA12) * mandrews@dcr.net * http://www.bit0.com/ VP, sysadmin, & network guy, Digital Crescent Inc, Frankfort KY Dialup/ADSL/ISDN/T1 Internet access for Frankfort KY and surrounding counties www.fark.com: If it's not news, it's Fark. (Or something like that.) On Fri, 15 Sep 2000, Kevin Hemsley wrote:
We currently use a variety of methods to monitor service quality, including SNMP polls and analysis of call statistics, MRTG graphing, etc. In addition to these, I would like to find a tool to repetitively dial POPs to test for various things: Ring-No-Answers, TCP/IP connectivity, Busy signals, etc. Although rare, we have had cases where customers have found significant problems that we had not yet identified. Problems that I feel we should have been able to identify faster than we did. I was about to embark on a Visual C++ project to build a tool to test POPs and alert us of anomalies. Before I do, I wondered if anyone has seen software that already performs these types of tasks?
Ultimately, it would be nice to have users with a dialer that automatically reported dial-in statistics back to a server each time it successfully connected. This would provide real-time statistics from potentially thousands of "sensors". The problem is that I haven't found a product that has a price tag that we would even consider. Is anyone using anything like this?
---------------------------------- Kevin Hemsley Systems Engineer Microserv Computer Technologies, Inc. kev@ida.net KB7TYA
- To unsubscribe to usr-tc, send an email to "majordomo@xmission.com" with "unsubscribe usr-tc" in the body of the message. For information on digests or retrieving files and old messages send "help" to the same address. Do not use quotes in your message.
- To unsubscribe to usr-tc, send an email to "majordomo@xmission.com" with "unsubscribe usr-tc" in the body of the message. For information on digests or retrieving files and old messages send "help" to the same address. Do not use quotes in your message.