Radical security development rethink urged

Traditional technology management methods are not capable of dealing with development of mission-critical security projects effectively, an industry firm claimed today.

Benjamin Jun, vice president of Cryptography Research, warned that fundamentally different engineering processes are needed because traditional methods have meant managers can create systems that are impossible to check and difficult to repair.

"Designers working on mission-critical security must focus on minimising the odds of security flaws, yet most engineers have had little, if any, training to teach them the skills necessary to do this effectively," said Jun.

"Although it's impossible to prove that a system will be secure, many managers fail to take even simple steps towards reducing the possibility that they'll ship a disaster.

"By failing to consider how engineering choices affect the system's life cycle, managers can create systems that are impossible to validate and impractical to repair."

Jun believes the security technology industry is a field still in relative infancy, relating security development processes with more mature high-risk fields, such as nuclear power plant operations and surgical medicine.

"While students at every medical school meet weekly to discuss surgical errors and methods for preventing them, I don't know of a single computer science programme that meets even once a semester to discuss software bugs or practical ways to avoid them," Jun said.