Methodology for network security design

A possible methodology for network security design is investigated. The ideas for each phase of the method-specification, design, and implementation-are described in detail. An attempt is made to apply the method to an invented example. Several pitfalls that await the requirements specifier are discovered. One problem is that defining and classifying security services is not as straightforward as one would like. A more fundamental defect of the methodology is its rigid sequencing of specification followed by design followed by implementation. Moreover, it ignores the newer developments in the computing world, i.e. object-oriented programming and client-server computing.<>