SGER/ SMA: Qualitative and Quantitative Models of Redundancy

Project Summary

SGER/CSR-SMA: Qualitative and Quantitative Models of Redundancy

A. Mili, NJIT

The project will conduct research to explore the following aspects with respect to redundancy: 1) Study redundancy as a system attribute in its own right, alongside such system attributes as modularity, design integrity, complexity, size, etc. While most applications of redundancy that we can envision today are in fault tolerance, it may be worthwhile to characterize it independently of fault tolerance, if we are ever going to find whether redundancy is relevant to any other system behavior or system property; and 2) Explore formal models of redundancy, with the intent of perhaps producing a formal definition for it, and a formal basis for reasoning about it. Specifically, quantitative and qualitative models of system redundancy, that have been explored so far will be further pursued to derive a comprehensive model that captures all aspects and all sources of a redundancy in a complex system.

The context of this one-year project, is to pursue exploration of the qualitative and quantitative models of redundancy, with the following goals in mind: a) Turn the individual qualitative observations into unified comprehensive model that characterizes redundancy as a system property reflecting the attributes of its state representations, system functions, and system specifications (all of which are sources of redundancy); b) Unify the qualitative and quantitative model, essentially by making the quantitative model measure the attributes that are captured in the qualitative model. We liken the contrast between the qualitative and quantitative model to characterizing a set (qualitative) and computing its cardinality (quantitative); whereas the former is a faithful reflection of the set, the latter is a convenient abstraction; c) Use the derived models to explore new insights and application; and d) Draw a longer term research plan, to the extent warranted by our preliminary results.