posted on 2009-05-14, 08:38authored bySergiy A. Vilkomir, David Lorge Parnas, Veena B. Mendiratta, Eamonn Murphy
This paper presents a method of estimating the availability of fault-tolerant computer systems with several recovery procedures. A segregated failures model has been proposed recently for this purpose. This paper provides further analysis and extension of this model. The segregated failures model is compared with a Markov chain model and is extended for the
situation when the coverage factor is unknown and
failure escalation rates must be used instead. This situation is illustrated in detail by estimating availability of a Lucent Technologies Reliable Clustered Computing architecture. For this example, numeric values are provided for availability indexes and the contribution of each recovery procedure to total system availability is analysed.
History
Publication
Proceedings of the 29th Australasian Computer Science Conference (ACSC 2006);