Confidentiality and Information Security

When it comes to confidentiality, a good information security strategy is to adopt the need-to-know basis for determining who has access to which data and when they have access to those data. Essentially, this paradigm states that a user should, by default, have access to no system capabilities or information assets. The assets or capabilities that are ultimately granted to the user, then, are done so only on a need-to-know basis. Similar to the need-to-know policy for data access, access to physical assets such as a server room or a network closet should also be granted only on a need-to-know basis. Put another way, system users and information technology workers should be provided with all of the information assets and access that they need to do their jobs effectively, and nothing more.

Another interesting consideration with respect to confidentiality is the question of how we know if a user truly is the person or system that they claim to be. This question speaks directly to the difference between identification and authentication. In generic terms, identification can be thought of as the process of proving that someone is who they say they are. By contrast, authentication is the process of proving that something is genuine, true, or authentic. In the world of information security, it is often very difficult or infeasible to truly identify a real human being or a specific system. Instead, we commonly use methods of authentication, and in so doing we assume that the credentials being used for purposes of authentication are being used only by the real-world system or human being to whom those credentials apply. This is, of course, a risky assumption, since through malicious or non-malicious means it might be very possible for another person to obtain your login credentials. If that person were then to use those credentials to login to, say, your social networking account, as far as the social networking site is concerned, that person is you. After having received appropriate credentials, the system will assume that the malicious party is, in fact, the real-world human being to whom those credentials actually belong. Confidentiality is thus difficult to ensure with 100% certainty, but it is nevertheless often the easiest security goal to assess in terms of whether or not efforts aimed at ensuring confidentiality have been successful.