基于控制关系和数据关系的科学工作流验证

Scientific Workflow Verification Based on Control Relationship and Data Relationship

面向科学问题求解的科学工作流具有以数据为中心的特性,其验证要求同时考虑控制流和数据流的合理性。为此,分析科学工作流中的4种控制关系和2种数据关系,给出相关合理性定义,提出相应算法以遍历获得科学工作流中存在的控制关系和数据关系,实现对科学工作流的合理性验证。通过正反实例证明了该验证方法的有效性。

Scientific workflows oriented to solving processes of science problems are often data-centric.The verification of scientific workflows needs to take account of the rationality of both control-flow and data-flow.In this paper,four kinds of control relationships and two kinds of data relationships.After that,the rationality definition of scientific workflow is presented based on those relationships.Sequentially,relevant algorithms are proposed to obtain kinds of relationships among activities in scientific workflows and then the verification on scientific workflows could be realized according to those relationships.Several positive and negative examples are provided to illustrate the effect of this proposed approach.