Chewing Gum and Bailing Wire


  • Page 1 of 1
    Bookmark and Share

We in the embedded industry are used to dealing with issues of reliability and robustness. We tout ruggedized designs, ballyhoo “five nines” reliability and the ability to keep systems going with hot swap. We subject modules and systems to grueling shock and vibration tests. We fret over the specs of connectors. Why then in the name of common sense do we put up with a national electrical power grid that is held together with chewing gum and bailing wire? This thing is an antique and it is a national disgrace.

I have been to a number of conferences where I have sat through numerous presentations about how vulnerable the grid is to terrorist attack by malicious hackers. These are, of course, credible and scary enough, but evil hackers are not the only threat to the grid. There are also things like squirrels. There are causes that we have not figured out. There are also more exotic threats from things like solar storms that have caused power outages and that have the potential–in the case of a coronal mass ejection (CME)–to shut down the entire grid for a very long time. As I write this, almost the entire state of Kentucky is paralyzed by a massive outage caused by a snow storm–a snow storm!

As we become ever more dependent on the gadgets we all design and sell, and on the Internet to link us all together with each other and with these devices, it helps to remember that they all depend on the grid as a source of electricity. If it fails too broadly and for too long as the result of a catastrophic outage, we are all living in Amish country. How can it be that something so vital to the continuation of modern society can be so vulnerable? The much-touted vision of the “Smart Grid” promises all sorts of benefits like smart appliances, time-specific pricing and all that, but without basic reliability–which does not now exist–these are all secondary.

There is a cliché that says if Alexander Graham Bell were suddenly to arrive in the 21st century in a modified DeLorean, he would not recognize the current state of telephony–cell phones, texting, BlackBerrys, etc. But if Thomas Edison were to show up, he would be very familiar with the grid–its basic infrastructure, that is. There are some pretty dazzling things behind the scenes. That is partly because over the years, the grid has simply been playing a game of keeping up with electricity demand. It has been constantly expanded with the main goal of keeping the lights on. Efficiency, robustness, visibility, environmental impact, adaptability have all taken a minor roll. The Department of Energy has estimated that if today’s grid were to be made just 5% more efficient, the savings would be equivalent to eliminating the fuel and greenhouse gas emissions of 53 million vehicles.

To someone who is not initiated in the inner mysteries of the grid but who is quite familiar with technology issues in general, there are several disturbing impressions. One is that among grid engineers, operators and utilities there is a huge disinclination, verging on paranoia, about doing anything substantial to change the grid. It is almost as if the thing is so rickety they fear any attempts to improve it may lead to total collapse. Add to that the daunting hurdles of government regulation and certification and you approach near paralysis. At one conference, which was generally pretty optimistic and forward looking, the only presenter who talked about problems that seemed too big to tackle was a grid engineer. When asked about the coming age of renewable power sources coming online in decentralized locations, his only reaction was to downplay the idea because of the “system stability problems” they would cause. Needless to say, this guy was out of touch with the times. If the grid is that vulnerable, we can’t just ignore the problem by not connecting renewable sources. That won’t make the vulnerability go away.

In fact, in many cases, decentralization is a key to robustness. Today’s grid seems so connected and without any checks that a failure can propagate far and wide when there ought to be a way to contain it. Obviously, we still need to be able to move power around to meet shifting demands, but there also needs to be intelligence, the ability to implement isolation, and redundancy to have a more robust and reliable power distribution system. We certainly have the machine intelligence in the form of all sorts of embedded controllers, sensors and systems. What we don’t seem to have is a level of situational awareness and an architecture that can move to a more decentralized and hence more robust model. And until that changes, including enhanced security against intentional attack, society itself is at risk.