Modern automotive and aerospace products are large cyber-physical system consisting of software, mechanical, electrical and electronic components. The increasing complexity of such systems is a major concern as it impacts development time and effort, as well as, initial and operational costs. Although much literature exists on complexity metrics, very little work has been done in determining if metrics correlate with real world products. Aspects of complexity include the product structure, development process and manufacturing. Since all these aspects can be uniformly represented in the form of networks, we examine common network based complexity measures in this paper. Network metrics are grouped into three categories: size complexity, numeric complexity (degree of coupling) and technological complexity (solvability). Several empirical studies were undertaken to determine the efficacy of various metrics. One approach was to survey project engineers in an aerospace company to gauge their perception of complexity. The second was through case studies of alternative designs to perform equivalent functions. The third was to look at actual time, labor data from past projects. Data structures and fast algorithms for complexity calculations for large cyber physical systems were also implemented.