Adaptive Reasoning Algorithm with Automated Test Cases Generation and Test Algebra in Saas System
A new integrated testing framework is proposed to use adaptive reasoning algorithm with automated test cases generation (ARP ) and test algebra (TA ) for increasing SaaS testing efficiency in faulty combination identification and elimination. The ARP algo
- PDF / 1,108,677 Bytes
- 17 Pages / 439.37 x 666.142 pts Page_size
- 60 Downloads / 196 Views
Adaptive Reasoning Algorithm with Automated Test Cases Generation and Test Algebra in Saas System
Abstract A new integrated testing framework is proposed to use adaptive reasoning algorithm with automated test cases generation (ARP) and test algebra (TA) for increasing SaaS testing efficiency in faulty combination identification and elimination. The ARP algorithm has been evaluated by both simulation and real experimentation using a MTA SaaS sample running on GAE (Google App Engine). Both the simulation and experiment show that the ARP algorithm can identify those faulty combinations rapidly and TA can eliminate a large number of faults from candidate test set with a small number of seeded faults.
7.1 Experimentation Using a MTA SaaS Sample This adaptive reasoning algorithm is evaluated using a sample MTA SaaS, and the SaaS was developed using the OIC framework, including customization of GUI, workflows, services, and data. The SaaS implements Arcade Game Maker (AGM) that tenant can configure their game applications and run on GAE. The games originated from software product line (SPL) system developed at the SEI/CMU. The overall SaaS architecture is shown in Fig. 7.1. On the left side, the system handles the access control as each role has its own access right. For example, an administrator is a role to manage all the customization information of tenants, but tenants can access their own configurations only and each player can execute the game only. The system has 15 components and each component has different number of configurable values as shown in Fig. 7.2. Assuming each component in the SaaS database has previously module tested, and thus faults in the SaaS system are those t-way component interaction faults. The number of possible configurations is 2,275,983,360. We have performed exhaustive
Parts of this chapter is reprinted from [1, 3], with permission from IEEE. © The Author(s) 2017 W. Tsai and G. Qi, Combinatorial Testing in Cloud Computing, SpringerBriefs in Computer Science, https://doi.org/10.1007/978-981-10-4481-6_7
83
84
7 Adaptive Reasoning Algorithm with Automated Test Cases Generation …
Fig. 7.1 Architecture of SUT
Fig. 7.2 All the configured components and elements
testing on these combinations to verify the SaaS system, and the faults are identified as follows: The conflict of objects’ borders: For any two objects, their starting positions may be overlapped. These kinds of faults are defined by four components, Object Xposition, Object Yposition, Object Width, and Object Height. The conflict of objects’ borders and boundary of game screen: For any objects, their starting positions might be out of game screen. These kinds of faults is defined by six components, Object Xposition, Object Yposition, Object Width, Object Height, Screen Width, and Screen Height.
7.1 Experimentation Using a MTA SaaS Sample
85
The conflict of color between objects and game screen: The object color cannot be the same as background color of the game board. These kinds of faults are defined by two components,
Data Loading...
