« Zooooooooom | Main | The Future is Plastic »

Monoculturalism

Diversity is good, particularly when it comes to networked computers. This was the conclusion of a report released in September at the Computer and Communication Industry Association meeting in Washington D.C., and is the subject of a new study funded by the National Science Foundation (and to be carried out jointly by Carnegie Mellon and the University of New Mexico), looking into ways to protect against viruses, intrusions, and other digital menaces.

If we apply the lessons of biology to computer networks -- a sensible approach, given that both have characteristics of complex adaptive systems -- the notion that diverse environments are more survivable than monocultural one makes a great deal of sense. A given bit of hostile code can't spread to every member of a network if the network contains a variety of different operating systems, just as a given tree disease can't spread to every tree in a stand if the forest contains a variety of different species. (Disclaimer/hype: I wrote about this very subject four and a half years ago, in Salon magazine.)

The lesson here for developers of networked, collaborative systems is to be open to diversity. A distributed network which allows varieties of devices (or operating systems) to participate can be more resilient when it is, inevitably, attacked. You may well be able to shrug off an attack entirely. Your members will thank you for your foresight.

Comments (2)

sinchlep:

I was working with a friend on a database for patients in his research on psychiatry and genetics. They used a form in doing their diagnosis but I designed the database for any form (basic approach in writing software). He said that is hardly ever changed but I did it anyway. Sure enough, for logistics reasons they had to adopt a totally different form. He said, "Glad one of us was thinking ahead, because we never do."

Then I mentioned version control to him another time. What's that?

Then it kinda hit me that the way to look at DNA is not as a data file, but as a self-perpetuating software system in a ZIP file (the DNA literally unzips), complete with compiler, data and version control.

This is complete speculation based on no knowledge, but it bugged me that people described "junk DNA" as junk. That can't be right, I thought because it is just too "expensive" to keep and can lead to errors. Just like leaving old code in files. But solutions to problems are expensive to generate, so you want to keep them. There may be a mechanism in our genetics to extract the old DNA (our forefathers) that are in the and mitocondrian.

Also, men, because they get both the X and the Y genes act as testers for code changes. If it kills a guy, no big deal, but if it provides an advantage then can be spread quickly. Women as the stable XX provide for long term stability while men can fluctuate more and act as a buffer.

Crazy talk, I know.

Accumulation of mutations is expected in non-functional genetic material. This shows also why diversity is not to be assumed to be valuable; functional genetic or other programs can't be randomized by diversity or they will suffer loss of function. In this sense monoculturalism is what every system strives for so long as it is functional; the pro-diversity even tries for it, even though it is self-contradictory in that instance.

About

This page contains a single entry from the blog posted on November 26, 2003 4:25 PM.

The previous post in this blog was Zooooooooom.

The next post in this blog is The Future is Plastic.

Many more can be found on the main index page or by looking through the archives.

Powered by
Movable Type 3.34