August 14, 2013
Boy, you just never know where the next apocalyptic data center event is going to come from. Since we are all well versed in the potentially adverse affects of hurricanes, tornadoes and earthquakes, I think that we are all pretty comfortable in understanding the inherent strengths and weaknesses of our facilities as they relate to adverse terrestrial activity. Not unlike the Three Little Pigs, we tend to build our data centers to withstand varying degrees of potentially catastrophic activity. Those companies that are located in areas that are less subject to wild deviations in atmospheric and seismic activity may elect to build their data centers in a fashion analogous to the middle pig’s house of sticks, while those in more tempestuous locations can elect to take the “house of bricks” approach. The main point here is that we all know the conventional risks and make our decisions accordingly. So, just when I think I’ve come up with a tidy little analogy—who didn’t like the Three Little Pigs when they were a kid, and was there a better villain than the Big Bad Wolf?–to describe the data center design continuum, I read an article warning of the potential for some nasty “space weather”.
That’s right kids, the next big threat to your data center might not be that hurricane gathering steam somewhere in the Atlantic, but a humongous—excuse the technical terminology—solar flare. Apparently the epicenter of our solar system is in the midst of switching magnetic polarities and that increased level of activity raises the potential for the sun to hurl a boatload of solar particles and gamma rays our way at any time. Some of you may feel that since we have a nice 93 million mile buffer that such an event might wreak havoc on Mars and Venus while leaving us unscathed, but you would be wrong. Apparently the heliosphere extends billions of miles past Pluto, so even if your DR plan is to build a back-up facility on Neptune, you’d still be out of luck.
For those of you glass half empty folks who are wondering how bad a solar assault could be; a quick look at the worst solar event in our history can provide some perspective. The “Carrington Event”, named after the guy that put two and two together and identified a large solar flare as the progenitor of strange occurrences here at home, happened in 1859. Among the resulting phenomena were the northern lights being seen as far south as Cuba and telegraph operators reporting sparks leaping from their equipment. If a similar event were to occur today, it is estimated that we could incur months long blackouts—particularly along the densely packed eastern seaboard—and thousands of deaths. Satellite systems would be particularly affected, and as a DirecTV subscriber, I think that I can speak for all of us when I say that this devastation would probably not be so bad if it happened after the coming football season, but apparently these things tend to be rather unpredictable.
While these dire potentialities definitely beg the question of how best to prepare for such an event, there appears to be no real panacea. Other than a robust back-up structure with UPS systems, a standby generator or two and a good pair of Raybans there is no data center equivalent to an SPF 100 sun block to ward off the ill effects of the unexpected solar lash. The good news is that the potential for such a devastating attack emanating from our sun is highly unlikely. After all, it has been 154 years since the infamous Carrington Event. Fortunately, the recommended prescription for protecting your data center from a solar event are really just the same basic practices that we should all be implementing within our facilities as part of the house of bricks approach. While this may seem like overkill to some of you out there, the rest of us will view this potential for solar malfeasance as just another example of that old data center maxim that, “danger lurks behind every corner”.