Abstract
The paper presents a survey of out-of-core methods available for the analysis of large Markov chains on single workstations. First, we discuss the main sparse matrix storage schemes and review iterative methods for the solution of systems of linear equations typically used in disk-based methods. Next, various out-of-core approaches for the steady state solution of CTMCs are described. In this context, serial out-of-core algorithms are outlined and analysed with the help of their implementations. A comparison of time and memory requirements for typical benchmark models is given.