Lifting Low-Level Workflow Changes Through User-Defined Graph-Rule-Based Patterns

被引：2

作者：

Jahl, Alexander ^{[1
]}

Baraki, Harun ^{[1
]}

Tran, Huu Tam ^{[1
]}

Kuppili, Ramaprasad ^{[1
]}

Geihs, Kurt ^{[1
]}

机构：

[1] Univ Kassel, Distributed Syst Grp, Kassel, Germany

来源：

DISTRIBUTED APPLICATIONS AND INTEROPERABLE SYSTEMS, DAIS 2017 | 2017年 / 10320卷

关键词：

Graph transformation; Graph matching; Pattern matching; Change Impact Analysis; Dependency graph; Web services; Service evolution; Answer set programming;

D O I：

10.1007/978-3-319-59665-5_8

中图分类号：

TP3 [计算技术、计算机技术];

学科分类号：

0812 ;

摘要：

In dynamic service-oriented architectures, services and service compositions underlie constant evolution that may not only affect the own workflow but dependent services too. Subsequently, required adaptations necessitate an effective detection of the changes and their effects. Merely capturing a sequence of low-level changes and analyzing each of them demands much coordination and may lead to an incomplete picture. An abstraction that summarizes a combination of low-level changes will facilitate the detection and reduce the number of changes that shall be considered for adaptation. In this paper, we propose an abstraction that is formulated through graph-based patterns, since service compositions are workflows that can be mapped to directed labeled graphs. The characteristics and granularity of a graph pattern can be adjusted by domain experts to the respective workflow language and application case. In particular, graph-based patterns are crucial when workflows are represented in two different formats. This could be the case if there exists one representation for the execution and one for the verification. We present implementation details and a detailed example that shows the feasibility and simplicity of our solution.

引用

页码：115 / 128

页数：14