How do you monitor your system?
Because our system is small and run entirely by volunteers, we do not have a
staffed 24x7 network operations center. The only system monitoring that we do
is automatic. Monitoring provisions include:
- What's Up. We use
What's Up
Gold from Ipswitch. It pings each of our important network nodes, both
upstream to our ISP (and to its ISP) and downstream to routers in each of our
subscribers' homes. In this way we monitor:
- our ISP's connection to the Internet backbone
- our frame relay T1 line
- our point-to-point microwave link from the T1 line to our neighborhood
- our DSLAM
- the DSL links to each of our subscribers (how we do
it)
- the IP connectivity to the DSL routers of the subscribers
If What's Up detects a failure to respond, it logs the event and sends a text
page to predetermined destinations such as cell phones and email addresses of
system volunteers.
- MRTG. We use Multi-Router Traffic
Grapher to graph traffic through our system. We separately graph the
traffic on a number of links including:
We inspect the graphs regularly to see if any of our links comes close to being
full. In this way we would know if we needed to increase capacity, for example
on our feed to the Internet backbone. You can see excerpts
from our MRTG configuration file.
- UPS self-tests. Our uninterruptible power supply tests itself every
two weeks and emails the results of the test to system volunteers. If the UPS
batteries were to require replacement we would thus receive early warning.
- DSLAM and UPS environmental monitoring. We monitor the UPS and DSLAM
environment. You can see temperature and
humidity and line voltage
trends. Temperature readings are taken inside the UPS (UPS battery temperature)
and in the cabinet that contains the DSLAM and UPS (cabinet ambient
temperature).
- Microwave link monitoring. The point-to-point microwave system
automatically steps down its data rate (from 3M bps down to 1M bps) if
conditions degrade the wireless connection. We
monitor and graph the data
rate.
- Security monitoring. The Dillon office of
Oppedahl & Larson LLP, which contains
a router linking the frame relay T1 line to one end of the point-to-point
microwave system, is alarmed with perimeter and motion sensors. The DSLAM
cabinet in the barn is also alarmed. Violation of either alarm results in text
pages to Coop volunteers.
- Qwest network operations. The provider of our frame relay T1 line
(Qwest) monitors the line continuously. If Qwest detects a loss of service, it
contacts us.
- What's Up Gold can receive SNMP traps, and our DSLAM can send SNMP trap
messages to annunciate system events. At some point we plan to configure our
DSLAM and our What's Up Gold to monitor and annunciate selected SNMP traps.