Consider a few personnel incidents that made the news in the last few years:
Nick Leeson, an investment trader at the Barings Bank office in Singapore, and Toshihide Iguchi of the Daiwa Bank office in New York City, each made risky investments and lost substantial amounts of their bank's funds. Rather than admit to the losses, each of them altered computer records and effectively gambled more money to recoup the losses. Eventually, both were discovered after each bank lost more than one billion dollars. As a result, Barings was forced into insolvency, and Daiwa may not be allowed to operate in the United States in the future.
In the U.S., agents and other individuals with high-security clearances at the CIA, the FBI and the Armed Forces (Aldrich Ames, Jonathon Pollard, Robert Hanson, and Robert Walker, to name a few) were discovered to have been passing classified information to Russia and to Israel. Despite several special controls for security, these individuals were able to commit damaging acts of espionage?in some cases, for more than a decade.
John Deutch, the director of the CIA under President Clinton, was found to have taken classified government information from the Agency to his house, where the information was stored on classified computers configured for unclassified use and appropriately marked as "unclassified." While the classified information was resident, these same computers were used to access pornographic web sites?web sites that could have launched attacks against the computers using both public and undisclosed security vulnerabilities. Yet despite the fact that numerous policies and laws were broken, no administrative action was taken against Deutch, and Deutch was issued a presidential pardon by Clinton on Clinton's last day in office.
If you examine these cases and the vast number of computer security violations committed over the past few decades, you will find one common characteristic: 100% of them were caused by people. Break-ins were caused by people. Computer viruses were written by people. Passwords were stolen by people.
Clearly, without people, we wouldn't have computer security problems! However, because we continue to have people involved with computers, we need to be concerned with personnel security.
"Personnel security" is everything involving employees: hiring them, training them, monitoring their behavior, and, sometimes, handling their departure. Statistics show that the most common perpetrators of significant computer crime in some contexts are those people who have legitimate access now, or who have recently had access; some studies show that over 80% of incidents are caused by these individuals. Thus, managing personnel with privileged access is an important part of a good security plan.
People are involved in computer security problems in two ways. Some people unwittingly aid in the commission of security incidents by failing to follow proper procedures, by forgetting security considerations, and by not understanding what they are doing. Other people knowingly violate controls and procedures to cause or aid an incident. As we have noted earlier, the people who knowingly contribute to your security problems are most often your own users (or recent users): they are the ones who know the controls, and know what information of value may be present.
You are likely to encounter both kinds of individuals in the course of administering a Unix system. The controls and mechanisms involved in personnel security are many and varied. Discussions of all of them could fill an entire book, so we'll simply summarize some of the major considerations.