Now the digital world has a parallel in a germ of its own: the computer virus. These rogue programs loom large in the public imagination as computer cannibals, bits of code released by mischievous or vengeful hackers that will attach to programs and eat data and discs alike. But viruses don’t have to be bad. A few scientists have begun to argue that viruses are actually living organisms, capable someday of evolving into autonomous Net-runners that will retrieve information for their owners.
One of the leading exponents of the virus as-your-servant theory is Fred Cohen, a computer scientist and consultant. His 1984 Ph.D. thesis started the field of virus-security research. “A lot of the new technology that’s being used to update networks is actually virus technology,” he says, “but they don’t use the word because it has other, negative connotations.” In his book “It’s Alive,” Cohen describes networks into which he’s released LPs, or live programs, to perform several different types of housekeeping tasks. They distribute software and clean out old files. Cohen suggests that LPs could even gather information from multiple databases.
In the most radical view, viruses are an artificial life form. They live, breed, evolve, procreate and die. Genetic algorithms–segments of programs that allow computer code to change its own form over a reproductive iteration–are not new to computing. But such algorithms can be paired with another program called a Reaper that kills older and damaged programs. Put the two together and you get a pretty good model of evolution.
That’s how a kind of virus “game preserve” called Tierra works. Developed by Tom Ray, a University of Delaware ecologist, Tierra provides an environment in which self-replicating programs–viruses–compete for resources and evolve. Unlike the real world, inside Tierra tiny changes in a program don’t usually result in terminal errors. This tolerance lets the programs evolve in unpredictable ways, such as the development of smaller parasites that reproduce by using other viruses’ programming. That electronic biodiversity led Ray to propose seeding the entire Internet with progenitor viruses. He hopes that when faced with the richness of the Net, they’ll become more complex. Eventually they’d evolve into useful applications, and users of the network could domesticate them.
These utopian visions aren’t likely to come about any time soon. In fact, viruses of the type found in “the wild,” on working computers, are becoming of a threat. Right now their primary mode of transmission is still the diskette. Infections move and outbreaks cluster around their points of for the first few months. That gives researchers a chance to parse the viruses and defeat them.
But the connectivity of computers on networks–such as the Internet–increases the danger. Jeff Kephart and fellow virus experts at IBM’s Watson Research Center are developing an “automated immune system,” software that captures viruses, analyzes them and spreads information on killing them along a network.
Ultimately, there’s the viral Frankenstein issue: if they’re alive, will viruses outrun human control? Consider the trajectory of the first viruses, the “worms” at the famed Xerox Palo Alto Research Center of the 1970s. The worms worked PARC’s computer network, fetching messages and running diagnostic tests. One night a PARC worm ran amok, madly duplicating itself, finally crashing the system. PARC shut down the network until a software vaccine was developed–a matter of hours. But what happens when these critters start evolving faster than vaccines can be written?
