The bigger they are, the harder they fall and maximum size can bring maximum failure
I launched a BSD-oriented hardware company this week and coincidentally watched the documentary Food, Inc. along with reading the book "Everything is Going to Kill Everybody." Here are a few of the things I discovered about monocultures and why we should learn to recognize them.
Food, Inc. documents the hyper-mechanization of farming in a quest to synthesize food in the complete absence of farms, farmers, farmworkers or livestock. While this goal has yet to be achieved, the food industry is already acting as if it can produce a chicken breast in the same linear way the computer industry produces semiconductors. They're finding however that this approach can bring a unexpected challenges: the hyper-production of corn has lead to the practice of feeding corn to cows in service of the hyper-production of beef, which in turn leads to periodic E. coli bacteria outbreaks. These outbreaks in turn result in periodic deaths of the consumers of the resulting beef products. Seeing this biological fact as a bug in the system, literally and figuratively, the industry has "solved" it by bleaching beef products in the same manner they bleach textile and paper products.
Search "pink slime" for details.
Chapter two of "Everything is Going to Kill Everybody" describes Klebsiella planticola, a genetically-modified bacteria that was intended solve the problem of farmers burning fields after the harvest: the bacteria would simply turn the remaining plant matter into an alcohol suitable for various uses. Clean room testing by scientists determined that this goal was achievable, but additional independent testing found that it also ran the risk of continuing to modify all living plant matter it came in contact with.
rm -rf plantae*
Parallel to the scientific concerns raised by the Klebsiella planticola story is the fact that the investigating scientist, Dr. Elaine Ingham was swiftly discredited by the genetic engineering industry and that most accounts of the incident are rather ambiguous. The incident is not mentioned on Wikipedia despite the fact that it was never formally refuted.
What we see are the tempting opportunities and catastrophic risks of monocultures. In computing, Internet-borne worms and viruses such as Conficker demonstrate these risks by swiftly infecting and disabling or repurposing millions of computers at once. Where monocultures were once virtually impossible for societies to achieve, it is with some irony that I am communicating with you using English and TCP/IP. Such standards are indeed remarkably useful but also also provide attack vectors against all participants, without exception. Biological and technological outages are part of life but for the first time, they are possible on a global scale. We need look no further than Twitter's outage last week as an example of just how many people these technological outages can affect.
Out like a light switch. Think about that.
Herein lies the single greatest opportunity for BSD: the BSD's are the only proven nonproprietary alternatives to GNU's Not Unix/Linux. With Google, Facebook, Twitter and most supercomputer data centers relying on GNU/Linux, GNU/Linux-targeting malware could cause virtually-instant outages around the world just has Windows-targetding malware has in the past. This is by no means to say that the BSDs are impervious to attack, but that shareholder-obliged CTOs at a monoculture-aspiring companies should be fully aware of what monocultures they depend on because all monocultures run the risk of total failure.
Copyright © 2011 – 2014 Michael Dexter unless specified otherwise. Feedback and corrections welcome.