Sooner or later—by intent or by accident—we will face a catastrophic breakdown of the Internet. Yet we have no Plan B in place to reboot a rudimentary, low-bandwidth emergency communication network if the high-bandwidth system we have come to depend on fails.
In the event of a major network disruption, most of us will have no idea what to do except to try and check the Internet for advice. As the system begins to recover, the resulting overload may bring that recovery to a halt.
The ancestor of the Internet was the store-and-forward punched-paper-tape telegraph network. This low-bandwidth, high-latency system was sufficient to convey important messages, like "Send ammunition" or "Arriving New York Dec. 12. Much love. Stop."
We need a low-bandwidth, high-latency store-and-forward message system that can run in emergency mode on an ad-hoc network assembled from mobile phones and laptop computers even if the main networks fail. We should keep this system on standby, and periodically exercise it, along with a network of volunteers trained in network first aid the way we train lifeguards and babysitters in CPR. These first responders, like the amateur radio operators who restore communications after natural disasters, would prioritize essential communications, begin the process of recovery, and relay instructions as to what to do next.
Most computers—from you car's engine controller to your desktop—can be rebooted into safe mode to get you home. But no safe mode for the Internet? We should be worried about that.