A Research On Legacy System Business Rules Extraction Based On DC Slicing

With the continuous development of computer technology, especially the development of Web-based information dissemination and retrieval technology results in the evolution from the client/server(hereinafter referred to the C/S)architecture to a flexible distributed architecture. This evolution gives new meaning to the architecture in today's Web technology as the core application of the information network. This new type of distributed architecture is service-oriented architecture (hereinafter referred SOA) in which system functions consist of all kinds of undependable and reusable services. Compared with C/S architecture, SOA has a lot of advantages. For example, it is based on standard of industry. It is loosely coupled, coarse-grained, expandable, and flexible. And it can share services and its maintenance costs are greatly reduced.At present, there are a large number of C/S systems, which are the core of enterprise. After a long period operation, these systems have accumulated large amounts of data and documents which belong to these systems are already outdated. How to migrate these systems for SOA architecture, in order to reduce maintenance costs, has become a software engineering problem to be solved. That the key to solve the problem is how to extract business rules and make them become SOA Architecture services. This paper is an in-depth study.This paper will complete the following aspects:1. The framework of business rules extraction will be improved and refined from the following aspects:(1)Based on the concept of the graph, a Call Graph and a File Containing Graph will be defined.(2)A File Containing Graph generation algorithm will be developed to allows users to understand relations between files in the whole system.(3)A Call Graph generation algorithm will be developed to generate system slicing.(4) In order to extract the business rules of large-scale systems, a system slicing algorithm will be developed.(5)In order to obtain more accurate business rules, a domain knowledge identification algorithm will be developed.(6)By analysis and comparison of existing program slicing schema, concluded that the DC slicing schema has advantages on slicing cost and slicing precision. As a result, dependence-cache (hereinafter referred to DC) slices technology will be introduced to extract the business rules.(7)XML is used to represent business rules internally in order to improve extensibility and efficiency of business rules representation.(8)The principle of program slicing standards changes will be put forward in order to provide guidance to users.2. Visual C++ 6.0 development tool is used to develop a business rules extraction prototype system.3. Through test of the system, proved that the system has good efficiency in view of slicing precision.In this method, the paper provides some improvement and innovation mainly in the following four aspects.1. The DC slices technology will be introduced to extract the business rules of legacy systems.2. In order to extract the business rules of large-scale systems, calling relationship between modules will be introduced.3. In order to obtain more accurate business rules, domain knowledge will be introduced.4. The paper will abstract legacy systems into a simple directed graph of language independent.On the basis of the above, the author use Visual C++ 6.0 development technology to build a business rules extraction prototype system which is based on the C/S architecture and use the tool to extract business rules from the C++ language legacy systems. This will lay the foundation for migrating the C/S architecture to SOA.