International Journal of Advanced Computational Engineering and Networking, ISSN: 2320-2106,
Volume-3, Issue-8, Aug.-2015
SOFTWARE DEVELOPMENT, TESTING AND UTILITY THEORY IN INSURANCE 1
DINESH KUMAR SAINI, 2NEBRAS ALTAR
1 Associate Professor and Coordinator BIT Program Faculty of Computing and Information Technology, Sohar University Research Fellow and Adjunct Faculty, University of Queensland, Brisbane, Australia Email:
[email protected] 2
Abstract-The insurance industry is growing at a very fast pace. With increase in volume and categories such as life, health and general insurance, managing the entire process has become quite complex. In order to achieve efficient systems, it is quite necessary for the testing phase of the system to be robust and effective as well. Insurers now opt for technology in order to balance efficiency and growth spending where it counts the most. The paper below discusses the importance of software testing in this particular field. We also discuss various methodologies deployed during the process of testing a system based for insurance companies and insurers. We also discuss the coupling of software testing with quality improvement i.e. how software testing not only detects existing bugs and errors of the system but also improves the overall performance and design of the system. Keywords-Software Development Life Cycle, Testing, Quality Assurance, XML, Utility Theorem, Test cases
1.
I. INTRODUCTION Software development is a tedious and complex process. Each domain software follow the standard life cycle which is shown in figure one. Insurance sector is growing and whole process is automated and managed by software now. There is numerous software applications are available that help in calculating the premiums and other actuarial process. Being a broad industry, the insurance market has two kinds of companies: One being the specialized ones and the other companies being the more comprehensive ones offering a wider range of products from general insurance to health and life insurances. The accumulated knowledge in this domain has led to the understanding that software testing under this field includes some key points. They are:
2.
3.
Roles and Permissions: In order to make sure that the software system runs perfectly requires the developers to exactly understand the insurance procedure. The basic roles are: 1.1. Insurance agents submitting the insurance application 1.2. Employee within the insurance company who checks and decides whether to approve or decline the application 1.3. The administrator with the software rights. Information Interfaces: Any change of information in any of the offices i.e. the front, middle and the back should be reflected evenly without any error. Change of Numbers: New software usually involves change in insurance policy percentages, such as interests on provision fees. Therefore making data verification and validation extremely crucial.
Figure1: Software Development Stages Software Development, Testing And Utility Theory In Insurance 72
International Journal of Advanced Computational Engineering and Networking, ISSN: 2320-2106,
Volume-3, Issue-8, Aug.-2015
important role while calculating the policy maturity date and the amount associated with it. Step 2: The XML message is now used by other components to generate another XML file, a new one this time. Negative testing: Restrictions for most of the values of the nodes are present. The restrictions usually consist of examples such as number of characters, type of values and business rules etc. Now negative testing involves entering values breaking the restrictions just to find out whether correct error messages are displayed and no output XML files are generated. Correct Output Value: Now all the correct values within the restrictions are entered and the XML file now generated is scanned for errors. Step 3: The message displayed within the XML file matches to that of the one displayed in the front office, similar to the first step. This step involves the verification of the input and the output nodes. The above step sums up the methodology which involves XML verification of critical data. This methodology is easy to implement and involves minimalistic errors. Also, in order to maintain quality, correct testing of XML content and transfer process is required.
II. XML DATA VERIFICATION This is the core of insurance software testing. The insurance software can basically be divided into 3 working divisions: The front, middle and back office. The examination of front office involves testing the UI of the system, the system basically used by the agents and employees of the insurance company. The UI’s basic function is to help agents create nonredundant, efficient customer files where various insurances can be applied. The employees of the company further accept or reject the files forwarded by the agents; the most prominent way of transferring data over channels is through XML format.
Fig 1.0: Relation between software testing and Quality management.
Step 1: to assure quality all the information generated by the front office is accumulated into one XML file and the following steps are considered: The visible fields of the information gathered are put into the XML file with correct nodes and values. If any user in the front office fails to or forgets to fill information in any field then the default value of the field should appear in the XML file and importantly the files should not remain blank. For e.g. if the user forgets to enter the start date of the insurance policy then the system by default enters current date instead of leaving it blank. There may be cases when some of the nodes are visible only in XML format and not visible to the users in the front office, even these fields are entitled to be checked and verified and should not be left unchecked . For e.g. the start date of any application is mentioned in the front office but it is not necessary the life of the insurance policy be mentioned there, it might be mentioned in the XML file generated once the application is submitted. This field cannot remain un attended to as it plays an
Fig 2.0 Core Functional Framework of testing
Thorough understanding of data and the process involved is required in order to achieve data accuracy throughout the system. III. EXTRACT, TRANSFORM AND LOAD PARADIGM One of the common paradigms used for data testing in insurance field is extract, transform and load. Extract (Data Requests): requesting data often means data availability and retrieval from various sources such as web servers and data warehouses. It includes many variables such as data accuracy, availability etc. all of which need to be correct and exactly as required, if
Software Development, Testing And Utility Theory In Insurance 73
International Journal of Advanced Computational Engineering and Networking, ISSN: 2320-2106,
not then it could lead to serious malfunctioning of the whole system. Transform (data Calculation): Data might not always be useful when entered in the system, it needs to be processed such that it holds importance to both the client and the company such as converting the data from the first office into useful XML files.
Volume-3, Issue-8, Aug.-2015
Load (Results): this part of the paradigm is produced on demand. Most of the entities of any result are stored in partial form and are compiled when requested through specialized commands. These commands along with the results need to be tested making sure of their integrity.
IV. CHALLENGES FACED DURING INSURANCE SOFTWARE TESTING
Figure 3: TCS software company insurance offering
output. The above challenge can be managed By automated data verification testing, given the frequency of cross checks required.
In figure 3 it shown TCS one of the bigest IT capmany offering solutions for insurance sector. Busienss transformations, customer experience and operational efficiency is addressed in this company solution.
Difficulties encountered while implementing Test Automation: Tool Selection for Automation: Not all applications in insurance are alike and neither are all tools, different applications may require different tools. It is this choice of tools which is difficult as far as automation is considered. Unstable UI: Frequent change in Ui means multiple scripts ienclosed in a good architecture. Scripts in huge amounts are tough to maintain. Custom Controls: Using controls other than the default controls (drop down lists, menus etc.) calls for customization. Custom controls are fruitful during manual testing but as far as automation is considered, it is quite tough to
4.1. Understandability of Domain knowledge One application of a customer might not necessarily involve a single insurance policy, it might contain multiple policies. The workflow and their types may vastly vary which means different work paths need to be designated for every type. This involves heavy duty understandability of the whole process which at times is quite a nuisance because software testers are not experts on insurance policies. 4.2. Checking Huge Documents This does not prove to be a technical difficulty but requires a considerable amount of patience and work. Insurance applications usually involve a lot of paperwork. The information in the applications should be regularly cross-checked with the paper
Software Development, Testing And Utility Theory In Insurance 74
International Journal of Advanced Computational Engineering and Networking, ISSN: 2320-2106,
maintain and operate different custom controls for different tools. Software and Environment Configuration: During an automated testing procedure, if an unknown error occurs, the process may either abruptly halt or need manual intervention rendering the whole concept of automated testing useless. V. PERFORMANCE TESTING
Volume-3, Issue-8, Aug.-2015
a small investment from a user’s point of view specially in the field of insurance where a small change in number may result in very high losses for the company and their clients. REFERENCES [1]
[2]
The functionality of a product might be top notch but if the performance does not meet the requirements, the product fails in the eyes of the customer. Performance testing basically stimulates an environment which behaves in order to verify how the system responds under heavy usage. Insurance sector as many other fields requires performance testing in order to check whether the systems deployed within the organizations fulfill optimum performance criterion. Performance testing also looks in for the weak sectors; the testing is done separately for the front, middle and back office. Doing so helps developers and testers isolate weaker zones if present. A few of the types of performance testing that can be used in the insurance sector are: Load Test: Includes a multiple user test that simulates a heavy load on the system to check for errors and delays. Also used to test the maximum threshold of the system which helps calculate the maximum load the system can bear. Stress Test: In contrast with the load test, this determines the stress under which the system fails and also determines the factors and reasons for failure. Stability Test: This test uses time as a unit of measurement i.e. determines the maximum durations under which a system can function normally and stably. Also checks the system for memory leaks.
[3]
[4]
[5]
[6]
[7] [8]
[9]
[10]
[11]
[12]
[13] [14]
CONCLUSION Software testing in insurance fields is an absolute must and no longer an extended accessory. Performance needs to be at a higher level than it has ever been, making it possible for insurance companies to entertain more applications with minimalistic error and maximum efficiency. Most importantly, testing not only traces out errors and bugs but at the same time assures quality of the system deployed. It saves future hassles and improves overall maintainability of the software. Undoubtedly,
[15]
[16]
Ross J Anderson, Liability and Computer security: Nine Principles. In Dieter Gollmann, editor, Computer Security ( ESORICS’94), LNCS 875, 231-245, Berlin Heidelberg, 1994, Springer Verlag Ross J Anderson, Why information security is hard-an economic perspective, 2001.http://www.cl.cam.ac.uk/~rja14/econsec.html Huseyin Cavusoglu, Birendra Mishra and Srinivasan Raghunathan, The effect of internet security breach announcements on market value: Capital market reactions for breached firms and internet security developers. International journal of Electronic Commerce, 9, 2004,69-104 Lawrence A. Gordon, Martin P.Loeb, and Tashfeen Sohail, A framework for using insurance for cyber risk management. Communications of the ACM, 46(3), 2003, 81-85, Hossack, B.I., Pollard, J. and Zehnwirth, B, Introduction to statistics with applications to general insurances.U.K. Cambridge University Press, 1983. Grzebiela, T. Insurability of electronic commerce risks. Proceedings of the 35th. Hawaii International Conference on System Sciences, 2002. Rainer Bohme. Cyber Insurance revisited. Workshop on the economics of information security (WEIS), 2005. G. Karsai, J. Sztipanovits, A. Ledeczi, T. Bapty. Modelintegrated development of embedded software. In Proceedings of the IEEE, (2003), pp. 145–164 Dinesh Kumar Saini and Nirmal Gupta “Class Level Test Case Generation in Object Oriented Software Testing, International Journal of Information Technology and Web Engineering, (IJITWE) Vol. 3, Issue 2, pp. 19-26 pages, march 2008. USA Dinesh Kumar Saini “Testing Polymorphism in Object Oriented Systems for improving software Quality” ACM SIGSOFT Volume 34 Number 2 March 2009, ISSN: 01635948, USA Lakshmi Sunil Prakash, Dinesh Kumar Saini and Kutti N.S. “Integrating EduLearn earning Content Management System (LCMS) with Cooperating Learning Object Repositories (LORs) in a Peer to Peer (P2P) architectural Framework” ACM SIGSOFT Volume 34 Number 3 May 2009, ISSN: 01635948, USA. Wail M.Omar, Dinesh K. Saini and Mustafa Hassan “Credibility Of Digital Content in a Healthcare Collaborative Community” Software Tools and Algorithms for Biological Systems in book series "Advances in Experimental Medicine and Biology, AEMB" Springer, Page No, 2010 Dinesh Kumar Saini “Sense the Future” Campus Volume 1Issue 11, Page No14-17, February 2011. Dinesh K Saini, Mustafa Hasan “Architecture and Classification algorithms for large P2P Digital Library” The Second International Conference on Networked Digital Technologies OpenConf system, Crez Republic Europe 2010. Dinesh Kumar Saini, Wail M. Omar “Software Testing For Semantic Service Oriented Architecture for E-Health Software Services” SERP'10 - 9th international Conference on Software Engineering Research and Practice (USA) http://www.worldacademy-of-science.org/, P.No. 240-246. Dinesh Kumar Saini and Hemraj Saini "Achieving Quality Through Testing Polymorphism in Object Oriented Systems,"3rd International Conference on Quality, Reliability and INFOCOM Technology (Trends and Future Directions), 24 December, 2006, Indian National Sciences and Academics, New Delhi (India). Conference proceeding.
Software Development, Testing And Utility Theory In Insurance 75
