23 08, 11 17:30
Today’s magnitude 5.8 earthquake in the Washington area came just as I was putting my thoughts together on seismic risks and telecom networks. This was in part inspired by my May trip to Japan, including Tokyo and Sendai, and discussions and observations while there. (During the Vietnam War my military service dealt with underground nuclear test detection so seismic issues are not entirely new for me.)
While we usual think about seismic threats as only applying to California and Alaska, so very large earthquakes have occurred elsewhere during the history of our country. These include the 1811 magnitude 7.7 New Madrid (Missouri) earthquake and the 1886 magnitude 7.2 Charleston (South Carolina) earthquakes. Thus there is a real risk of a major earthquake in lots of places although with a small probability.
I am concerned that the telecom industry and regulators have not connected this fact with the large scale phasein of fiber optics as the dominant long haul communications technology as well as a key factor in cellular backhaul. In the past, long distance microwave had a major role in this area and that is diminishing. Microwave systems can be readily made earthquake resistant by strengthening the towers that hold the antennas and the nearby equipment buildings.
The picture that has always been at the top of this blog is of an NTT Docomo tower near Sendai Station in the center of the recent earthquake. It was taken 3 years ago. At right is the same tower in May after the disastrous earthquake - no apparent damage. The Japanese Government and the Japanese telecom industry recognize that hardened microwave towers, such as this one in Sendai, are an important precaution in maintaining connectivity after major earthquakes and such towers are found throughout Japan.
Whereas microwaves pass through air that is unaffected by earthquakes, most fiber optic transmission lines are in buried cables that could be damaged by ground fissures associated with some large earthquakes. Now in the Japanese 3/11 earthquake such fissures were not a problem. There were 3 long distance fiber routes between Tokyo and Sendai. The one closest to the ocean was damaged by tsunami action, the other two were not damaged. But my Japanese friends are now reexamining the lessons of the 1995 Kobe earthquake. There was little or no impact on fiber at the time, but then again fiber was not in as common use as it is today.
In an earthquake that results in ground fissures, fiber optic lines miles apart could suffer related damage so an analytical assumption that they are independent would be wrong. Realistically, fiber facilities sometimes end up in the same conduit so independence of reliability can be more suspect.
During the Cold War the Pentagon took actions with FCC to influence network design first to improve reliability in a post nuclear attack scenario and later to inhibit microwave interception by the Soviets. The former was done under the AT&T monopoly, but the later was done after the onset of real competition for AT&T.
While such direct government intervention into network design might not be appropriate in this case, it appears that there is no hard data on the relative use of fiber versus microwave for long haul transmission (or cellular backhaul) in areas with a risk of large earthquakes. Perhaps industry and the appropriate federal agencies should open a dialogue on the topic.