A Static Change Impact Analysis Approach based on Metrics and Visualizations to Support the Evolution of Workflow Repositories

A Static Change Impact Analysis Approach based on Metrics and Visualizations to Support the Evolution of Workflow Repositories

Gustavo Ansaldi Oliva (Department of Computer Science, University of São Paulo (USP), São Paulo, Brazil), Marco Aurélio Gerosa (Department of Computer Science, University of São Paulo (USP), São Paulo, Brazil), Fabio Kon (Department of Computer Science, University of São Paulo (USP), São Paulo, Brazil), Virginia Smith (HP Software, Hewlett-Packard, Roseville, CA, USA) and Dejan Milojicic (Hewlett-Packard Labs, Hewlett-Packard, Palo Alto, CA, USA)
Copyright: © 2016 |Pages: 28
DOI: 10.4018/IJWSR.2016040105
OnDemand PDF Download:


In ever-changing business environments, organizations continuously refine their processes to benefit from and meet the constraints of new technology, new business rules, and new market requirements. Workflow management systems (WFMSs) support organizations in evolving their processes by providing them with technological mechanisms to design, enact, and monitor workflows. However, workflows repositories often grow and start to encompass a variety of interdependent workflows. Without appropriate tool support, keeping track of such interdependencies and staying aware of the impact of a change in a workflow schema becomes hard. Workflow designers are often blindsided by changes that end up inducing side- and ripple-effects. This poses threats to the reliability of the workflows and ultimately hampers the evolvability of the workflow repository as a whole. In this paper, the authors introduce a change impact analysis approach based on metrics and visualizations to support the evolution of workflow repositories. They implemented the approach and later integrated it as a module in the HP Operations Orchestration (HP OO) WFMS. The authors conducted an exploratory study in which they thoroughly analyzed the workflow repositories of 8 HP OO customers. They characterized the customer repositories from a change impact perspective and compared them against each other. The authors were able to spot the workflows with high change impact among thousands of workflows in each repository. They also found that while the out-of-the-box repository included in HP OO had 10 workflows with high change impact, customer repositories included 11 (+10%) to 35 (+250%) workflows with this same characteristic. This result indicates the extent to which customers should put additional effort in evolving their repositories. The authors' approach contributes to the body of knowledge on static workflow evolution and complements existing dynamic workflow evolution approaches. Their techniques also aim to help organizations build more flexible and reliable workflow repositories.
Article Preview


Large-scale workflow repositories, which may encompass thousands of workflows in real world settings, are intrinsically complex. Workflows in these repositories frequently link to each other, forming a complex network of dependencies. As workflows evolve, their number of elements and interconnections tend to increase. Furthermore, organizations often heavily rely on some of the out-of-the-box (OOTB) workflows provided by vendors. This means that modifying or replacing these core workflows can affect the large amount of other workflows that depend on them. Therefore, evolving workflow repositories poses a challenging task.

In this context, two problems may occur. First, workflow designers may become reluctant to apply changes to workflows. In this case, the repository becomes less flexible, since it neither leverages opportunities nor deals with the constraints of new technology, new market requirements, and new legislation (Casati, Ceri, Pernici, & Pozzi, 1998). Second, workflow designers may end up performing changes to workflows without knowing the associated impact, because it is too difficult to be aware of all interdependencies and evaluate how critical they are. In this case, the repository becomes less reliable, since inappropriate changes may induce side- and/or ripple-effects (Arnold, 1996). A side-effect is an error or other undesirable behavior that occurs as a result of a modification (Freedman & Weinberg, 1982). In turn, a ripple-effect occurs when a small change to a system affects many other parts of this same system (Stevens, Myers, & Constantine, 1974). In fact, previous research already showed that making software changes without visibility into their effects can lead to poor effort estimates, delays in release schedules, degraded software design, unreliable software products, and premature retirement of software systems (Mens & Demeyer, 2008; Souza & Redmiles, 2008; Swanson & Beath, 1989). In summary, by being less flexible and less reliable, the workflow repository also becomes less evolvable.

This paper reports the results of joint efforts from researchers and engineers from the University of São Paulo, HP Labs, and HP software in seeking innovative workflow evolution solutions to be integrated into the HP Operations Orchestration (HP OO) product. HP OO is a professional industry Workflow Management System (WFMS) that provides an OOTB workflow repository targeted to help organizations automate common IT operations. Customers can also leverage this repository to build their own custom workflows. Table 1 depicts HP OO common usage scenarios.

Table 1.
Common HP operations orchestration usage scenarios1

Driven by customers’ feedback, we decided to focus on enhancing HP OO’s change impact analysis features. Software change impact analysis concerns “identifying the potential consequences of a change, or estimating what needs to be modified to accomplish a change” (Arnold, 1996). The analysis aims to make the existing relationships among artifacts more explicit to humans, so that they can maintain and evolve software systems more easily. Change impact analysis information can then support planning changes, approving changes, accommodating certain types of changes, and tracing through the effects of changes (Arnold, 1996). Naturally, mitigating side- and ripple-effects have also been two commons goals of change impact analysis (Kagdi & Maletic, 2006).

Complete Article List

Search this Journal:
Open Access Articles
Volume 14: 4 Issues (2017)
Volume 13: 4 Issues (2016)
Volume 12: 4 Issues (2015)
Volume 11: 4 Issues (2014)
Volume 10: 4 Issues (2013)
Volume 9: 4 Issues (2012)
Volume 8: 4 Issues (2011)
Volume 7: 4 Issues (2010)
Volume 6: 4 Issues (2009)
Volume 5: 4 Issues (2008)
Volume 4: 4 Issues (2007)
Volume 3: 4 Issues (2006)
Volume 2: 4 Issues (2005)
Volume 1: 4 Issues (2004)
View Complete Journal Contents Listing