A graph pattern matching approach to software architecture recovery

Abstract

This paper presents a technique for recovering the high level design of legacy software systems based on pattern matching and user defined architectural patterns. Architectural patterns are represented using a description language that is mapped to an attributed relational graph and allows to specify the legacy system components and their data and control flow interactions. Such pattern descriptions are viewed as queries that are applied against an entity-relation graph that represents information extracted from the source code of the software system. A multi-phase branch and bound search algorithm with a forward checking mechanism controls the matching process of the two graphs by which, the query is satisfied and its variables are instantiated. An association based scoring mechanism is used to rank the alternative results generated by the matching process. Experimental results of applying the technique on the Xfig system are also presented.

Supporting Migration to Services using Software Architecture Reconstruction

“…The area of software architecture reconstruction has made substantial progress over the past several years [3,5,7,8,9,11,14,15,18,20,21]. A number of techniques and methods have been developed along with tools to support them [4,6,12,19].…”

Text mining biomedical literature for constructing gene regulatory networks

“…Among them, Apriori (Agrawal et al, 1994(Agrawal et al, , 1995 is one of the most efficient algorithms. In recent years, based on the idea of reducing normal graph mining to frequent item set mining, several such algorithms have appeared (Koyutürk et al, 2004;Kuramochi et al, 2004;Inokuchi et al, 2003;Sartipi et al,2001). But none of these algorithms show how to deal with inexact match graph mining.…”

Software Quality Attribute Analysis by Architecture Reconstruction (SQUA3RE)

“…Alborz is a user assisted reverse engineering tool designed for analyzing and recovering software architecture in the form of cohesive modules and subsystems (Sartipi & Kontogiannis 2001). The tool's operation is based on techniques from the areas of data mining, pattern matching and clustering.…”

“…Much research has been done on the methods and techniques to support ARE, resulting in a variety of approaches, such as manual reconstruction with tool support (Finnigan et al 1997), query languages for writing patterns that automatically build aggregations , clustering (Mendonça & Kramer 2001), data mining (Sartipi & Kontogiannis 2001), and the use of architecture description languages (Eixelsberger et al 1998). …”