The computer industry’s “next grand challenge” is the ever-increasing complexity of computing in the Internet era, with its global networks and proliferation of digital devices, says Paul M. Horn, a senior vice president who oversees the research labs at I.B.M., in a paper, “Autonomic computing.”

“Autonomic computing” is a biological metaphor suggesting a systemic approach to attaining a higher level of automation in computing.

Similarly, Mr. Horn says, the way to handle the complexity problem is to create computer systems and software that can respond to changes in the digital environment, so the systems can adapt, heal themselves and protect themselves. Only then, he adds, will the need be reduced for constant human maintenance, fixing and debugging of computer systems.

It will require research efforts over the next 5 to 10 years in areas like adaptive algorithms for software agents, self-healing server computers, artificial intelligence and control theory.