I’ve been doing a little reading today, since it feels like Friday around here, and came across an article about space storms possibly creating disaster situations over large swaths of the US. This is due to our heavy reliance on the power grid for, well, pretty much everything.
The second problem is the grid’s interdependence with the systems that support our lives: water and sewage treatment, supermarket delivery infrastructures, power station controls, financial markets and many others all rely on electricity… “It’s just the opposite of how we usually think of natural disasters,” says John Kappenman… “Usually the less developed regions of the world are most vulnerable, not the highly sophisticated technological regions.”
Taking this down a bit into the IT infrastructure, this reminds me how we can become dependent on our own infrastructure to do common or even uncommon tasks. Web interfaces in a power outage or misconfiguration will be down. Do you know how to expediently console into your devices? Can you work on a command line? Do you have the documentation on how your scripts operate so you could do it manually in an emergency? Could you interpret tcpdump output if your network is being crippled by a worm, preventing IDS use?
Some of this comes down to something I believe in: the simple fundamentals. Tools are great to make us more efficient, but at the end of the day good IT persons are not defined by their GUIs. They are defined much like good ol’ Unix tools: how well they can use the simplest building blocks to get their tasks done. And how they can creatively chain those simple tools together to do fabulous things.
This also goes into security. We are not defined by the automated tools we use (those that are are script kiddies), but rather whether we understand how those tools work and could emulate similar behavior using the basics if need be.
Further we can expand this into our virtual infrastructure. If the host goes down, or hell, even just your virtual center client box, are you dead in the water? Would you be able to stand up a (*shiver!*) physical web server quick and get critical apps working while the host is being operated on?
Finally, this does echo an aspect to one of the simple security maxims that I believe was quoted or made popular by Schneier or Geer: “Complex systems fail complexly.”