Research on the Scale-Free Feature of the Embedded Component System

Abstract

In order to improve the productivity and efficiency of the embedded component systems, we need better ways to characterizing their architecture and hidden properties. In this paper, we explore a new viewpoint that the embedded component systems can be studied as a large amount of components connected by the relationships. Here we name it the Embedded Component Network (ECN). Then we take the eCos for example and analyze basic statistics parameters. Results confirm that the eCos network exhibits typical characteristics of complex network. Then we introduce three classical node ranking algorithms, which have some shortcomings for the real eCos system. So we propose a new ranking algorithm, which provides a general framework for the classic ranking and reveals more position differences of eCos network topology. This result has some practical values to assist the maintenance and reusability of the embedded systems.