Resilient, Off-Grid, and Sustainable Computing

Overview

These days, it seems that most devices and services are built with Internet connectivity in mind, even when such connectivity makes no sense. For example, popular portable solar power stations — and even some gasoline generators — have companion applications that require Internet access to communicate with a cloud service. However, the entire purpose of this equipment is to provide backup electricity whenever the power is out. When the electrical grid is down, Internet connectivity is also frequently lost, degrading the backup device's capabilities (or even rendering it inoperable in the worst case).

A similar trend has been occurring with desktop computer systems, as they have evolved from standalone devices that could run from local disks to systems that depend on an Internet connection just to log in after booting. These computers might not be useful for much in the aftermath of a hurricane or natural disaster. Worse, a global event that significantly disrupts power generation and distribution could take out many of the routers and servers required to make the Internet function. The Carrington Event of 1859 gives us historical data to see that this risk is more than merely theoretical.

In a disaster recovery scenario, restoration of computing systems and the Internet will be our shared responsibility as computing professionals. As such, we must be careful not to wander into a chicken-and-egg situation in which we need a functioning Internet in order to be able to service the equipment required to bring the Internet back online. Even if a disaster isn't global in scale, we can still suffer significant local events that require rebuilding systems without the luxury of a quick Internet search. For example, it is virtually guaranteed that a major hurricane will affect our area at some future time, taking out power, Internet, and cell service.

To mitigate these issues, I am working to create resilient, off-grid, sustainable computing systems:

Among Internet communities, the cyberdeck concept can be an example of a system that meets all three of these criteria. Since cyberdecks tend to be one-off builds, they invariably have a range of capabilities and requirements. However, they do provide inspiration for a type of system that could be useful in a disaster scenario. Note that the system aspect is important here: we need to consider both the hardware and the software!

Since the Department of Computing Sciences is already using Raspberry Pi single board computers in a variety of classes, I'm designing initial systems around these devices. Projects will include constructing the physical systems as well as the software that will be run on them during an emergency event. For maximum synergy, I am planning to use the Private Desktop Environment projects to create the base software stack.

Available Projects