M5.0 in Southern Indiana - June 18, 2002

Here is a quick rundown of web traffic generated by the M5.0 event that occurred at 12:37 CDT on June 18, 2002. Overall, the Earthquake Hazards Program web servers performed well during the surge of traffic generated by this event.


Here is the traffic on the Earthquake Hazards Program web site, as reported by Akamai. The peak is about 230 hits/sec and just over 9Mbits/sec. The peak occurred at about 40 minutes after the event, which is probably due to the fact that this event was located in an area where people are not as aware of earthquakes as we in California are. California events always generate peak traffic at 10 minutes after the event, while people in the Midwest probably had to hunt a bit before they found our site. This also means that the traffic spike was spread out over more time. This can be seen in the graph, which is not nearly as sharply peaked as it is for an event in California. An example of this is the spike generated by the M4.9 near Gilroy on May 13. The smaller peak at about 14:00 probably corresponds to reports on the 5:00 and 6:00PM news in the Eastern and Central time zones.

The only problem we experienced during this surge was right around the time of peak traffic on the Program site. The site was observed to be a bit slow responding. Checking the two origin servers found that they had both hit their maximum of 150 Apache child processes. The maximum was increased to 600 on each server. After that, the number of processes peaked at about 350. The servers were not overloaded at this level, and response time was observed to be back to normal.


Here is the total hits and data rate for the Pasadena web site as reported by Akamai. Note that the hit rate peaks at about 70/sec and the data rate is about 13 Mbits/sec. Most of this traffic is due to people either viewing the map or submitting questionnaires for the Community Internet Intensity Map.


Here is a graph of the CIIM questionnaires submitted. Note that the peak rate is about 80/minute. This is well-within our tested capacity of 2400/minute. Note that the second, sharp peak at about 18:00 is due to a M3.0 event centered under the west side of Los Angeles.


Here is the traffic on the USGS Menlo Park site. This is one of the oldest earthquake web sites in the USGS, and traditionally has always had the most traffic. Note that the peak hit rate is less than 40/sec, and the data rate is only about 1.5 Mbits/sec. This is much less than the traffic on the Earthquake Hazards Program site, which indicates that the Program site is becoming better-known to the public.

The National Earthquake Information Center traditionally also experiences large surges in traffic after events. Here, the peak was about 118 hits/sec and 21 Mbit/sec. This is comparable to the traffic on the Program site, and also is an indication that the Progam site is becoming known to the public.



Stan Schwarz
Honeywell Technical Services
Southern California Seismic Network Contract
Pasadena, California
June 20, 2002