I first started having intermittent connection issues with my Time Warner Internet connection sometime in May of 2008. After dealing with a total of 15 different individuals in Time Warner Technical Support over the course of 8 months, I finally got the issue resolved on February 3rd, 2009. On February 3rd, a technician came out and replaced our modem for the second time. I’m not sure what made this change different, but our connection has behaved itself since. I posted the scripts I use to monitor my connection, reboot the modem, and parse the log file on my personal website a while back. Now that things appear to be working correctly, I decided to parse my logs and see what they show.
Jul 2008: 472
Aug 2008: 507
Sep 2008: 188
Oct 2008: 217
Nov 2008: 1084
Dec 2008: 935
Jan 2009: 348
Feb 2009: 160
Mar 2009: 144
Downtime is counted in minutes. That means that in February and March, we experienced around 5 minutes of downtime per day. Most likely these minutes were non-conservative and were not noticeable. During the worst months of November and December, we experienced more than a half hour of downtime each day. This was quite noticeable and often occurred in 3, 6, or 9 minute blocks of downtime.
The best conclusion to take away from this is that pinging a server every minute to check connection availability is not granular enough to be overly useful statistics, but it was very useful to ensure that our Internet connection did not remain down for hours at a time.
