Service-Based Enterprise Systems represents an emerging paradigm in software development. Over the years, its usage has increased exponentially in fields such as manufacturing, banking, telecommunication and healthcare due to its ability to promote reusability, loosely-coupled and scalability for software development. However to accommodate the evolution of business requirements in SBES, there needs to be in place a standard strategic approach. Moreover, handling business requirements evolution in natural language is very challenging when deluge scale of requirements are expected. Therefore, we propose an approach to automate the activity of evaluating the need to evolve phase which firstly conducted in service evolution life cycle. The proposed approach follows Change-oriented Service Life Cycle methodology as guideline to portray different phases in service evolution. As proof of concept we will select a few SBES projects to evaluate the feasibility of the proposed approach.
Making its first emergence last few years, Service-Based Enterprise Systems (SBES hereafter) had received a very wide attention from software developers around the globe. Gaining the popularity due to its high potential capability to promote and maximize reusability, scalability, interoperability and performance in software development [7].
Being recognized as an evolution of conventional enterprise system [2]; the
development process of a SBES is more less the same, which started with Requirements
Engineering (RE) Process. To the best our knowledge, there is inadequate of research
being conducted to standardize the application Requirement Engineering (RE) process
in SBES environment particularly in industrial settings. Muneera et al [3] highlighted
there are five most challenging issues of RE in SBES and one of them is Requirement
Change Management and Evolution. The challenge would be, to analyze the impacts
of change on the system. Moreover, it is difficult to modify the functionality existing
service especially when the service is being consumed by various users in different
context. On top of that, having the requirements presented in natural language would
a challenging task for the analyst to extract information and decide the need to evolve
for a service [
Usually, in an enterprise environment, the beginning of service evolution will start
with Chief Information Officer (CIO). CIO initiates IT goals that need to be fulfilled
by the enterprise. The Enterprise Architect is the person who will translate the goal
into meaningful Business Requirements. These Business Requirements, which usually
presented in natural language, will be forwarded to the Service Analyst who will be
conducting further analysis. The first part of this analysis is to evaluate the need
evolve. This activity is the first phase in Change-oriented Service Life Cycle
methodology as proposed by Papazoglou [4], which serves as guideline to handle deep
changes in service evolution. Generally, deep changes is defined as large scale,
transformational changes cascading beyond consumers of service possibly affected the
entire end-to-end service chain. It normally deals with re-alignment of business
process and policy induced [4]. Studies showed, there are numerous techniques,
framework or model [4], [7], [9] and [
We propose an automated approach to assist service analyst, conducting the evaluation of need to evolve activity in service evolution life cycle. 2
The core aim of this research is to promote an automated environment of handling business requirements evolution in Service-Based Enterprise Systems. 3
To achieve the specified aim, a few objectives are derived: (1) to investigate the current issues of business requirements in Service-Based Enterprise System context; (2) to review the current solutions related to business requirements evolution in ServiceBased Enterprise Systems (3) design and develop a strategic approach to handle business requirements evolution in Service-Based Enterprise Systems; (4) validate the approach with appropriate case study.
The main research question is formulated as follow: How to assess the effectiveness of automating the process in handling business requirements evolution under Service-Based Enterprise System context? The main research question is further divided as follow: (1) What are the current issues and problems in business requirements evolution under Service-Based Enterprise context? (2) What are the current available solutions to support business requirements evolution in Service-Based Enterprise Systems context? (3) How to overcome the problems and issues of business requirements evolution in Service-Based Enterprise Systems? (4) How to evaluate effectiveness the proposed solution? 5
To materialize this project, Design Science Methodology for Information Systems and
Software Engineering [
Based on the current findings, there is only one stakeholder involving for the study that is service analyst. As stated earlier, service analyst will be conducting series of tasks under evaluate need to evolve activity.
Upon receiving deluge scale of requirements, the service analyst wished to have faster analysis of requirements evolution with high precision of analyzing the evolution impact when conducting the need to evolve activity.
Evaluating need to evolve activity usually involves three sub-activities as depicted
in Figure 1. Classify causes of evolution is conducted to categorize types of evolution.
Taxonomy constructed by Zaiwen Feng [9], claimed there are three types of
motivations that lead to service evolution. Those motivations are recognized as perfective,
corrective and performance. Once the classification process is done, the service
analyst will assess scope it is extend due to the evolution. At this phase, service analyst
will assess how many existing requirements get affected with the evolution initiatives.
This activity is also known as inter-requirement change impact analysis [
To conduct the aforementioned activities, require huge amount of time. Dealing with natural language requirement is an ever-challenging task faced by the service analyst, as most of the requirements are presented in natural language format [linguistic]. Apart from that, to accurately measure the inter-requirement is also difficult and tedious work.
This research enhances the implementation of conducting need to evolve activity by automating the implementation. 5.2
To automate the classification process, researcher will employ the taxonomy
constructed by [9] combined with the classification algorithm using natural language
processing technique implemented by [
Once the classification activity is finished, the service analyst will have to assess any
affected requirement due to the evolution initiative. To automate the process, natural
language processing technique utilized in [
This activity requires the service analyst to collect existing service metrics to assist in
understanding the nature of services-in-scope and related services and provide a
baseline for comparative and determination of expected productivity, cost and service
level improvements [6]. We employ work done by [
The researcher will conduct an experiment to test the efficiency of proposed solution. A real SBES project will be selected with complete software engineering artifacts like requirements document in particular. 5.3
As for validation, the researcher will compare and contrast how close the result produced by the tools to those produced by service analyst. The evaluation is devised using precision and accurateness. In this validation, experts from the requirements engineering domain were used in this study.
Currently, the researcher has completed the literature study. Based on the findings from literature survey, a set of questions is constructed for the qualitative study. The qualitative study is conducted as an explanatory purpose to gain better understanding pertaining issues and challenges faced by the industrial practitioner. Conducting the interview session will also enrich researcher’s findings. Some of the undiscovered problems reported by the literature might be discovered during the interview sessions. Once the interview session is completed, the result will be assessed and analyze.
The next step to proceed would be the design of the treatment for the identified problem. This would take quite some times.
Acknowledgments. The PhD student is currently funded by Majlis Amanah Rakyat (MARA), a Malaysian Government Agency.