Philip Samuel

Automatic test case generation using unified modeling language (UML) state diagrams

UML is widely accepted and used by industry for modelling and design of software systems. A novel method to automatically generate test cases based on UML state models is presented. In the present approach, the control and data flow logic available in the UML state diagram to generate test data are exploited. The state machine graph is traversed and the conditional predicates on every transition are selected. Then these conditional predicates are transformed and function minimisation technique is applied to generate test cases. The present test data generation scheme is fully automatic and the generated test cases satisfy transition path coverage criteria. The generated test cases can be used to test class as well as cluster-level state-dependent behaviours.

Automatic test case generation from UML communication diagrams

We present a method to generate cluster level test cases based on UML communication diagrams. In our approach, we first construct a tree representation of communication diagrams. We then carry out a post-order traversal of the constructed tree for selecting conditional predicates from the communication diagram. We transform the conditional predicates on the communication diagram and apply function minimization technique to generate the test data. The generated test cases achieve message paths coverage as well as boundary coverage. We have implemented our technique and tested it on several example problems.

Slicing-based test case generation from UML activity diagrams

UML diagrams are important design and modeling artifacts. These diagrams can also be used to generate test cases. We present a novel test case generation method that is based on dynamic slicing of UML activity diagrams. We use flow dependence graph (FDG) of an activity diagram to generate dynamic slices. Dynamic slices are created using an edge marking method. Slices are constructed corresponding to each conditional predicate on activity edges and test cases are automatically generated with respect to each slice. Our generated test cases satisfy path coverage criterion. Our test data generation scheme is automatic and the test data is generated considering the slice condition. We have implemented our approach to realize a prototype tool.

Test Sequence Generation from UML Sequence Diagrams

In this paper, we present an approach to generate test sequences from UML 2.0 sequence diagrams. Sequence diagrams are one of the most widely used UML models in the software industry. Although sequence diagrams are used for modeling the dynamic aspects of the system, they can also be used for model based testing. Existing work does not encompass certain important features of UML 2.0 sequence diagrams. Our work considers many of the novel features of UML 2.0 sequence diagrams like alt, loop, opt and break to generate test sequences. These areimportant features as far as testing is concerned. Our work begins with defining the important types of relationship that can exist between the messages. Based on the relationship between the messages, the message sequences are generated. Our work considers an important feature of UML 2.0 sequence diagrams called the dasiaExecution Occurrencepsila to generate message sequences. Next, an intermediate representation of the sequence diagram is built. This intermediate representation is called the Sequence Dependency Graph (SDG). The message sequences are incorporated into the SDG. Finally, we discuss a traversal algorithm to generate test sequences from SDG. Our method is fully automated and the test sequences generated can be used to check the correctness of the implementation under test.

Web services regression test case prioritization

Web services and their underlying system grow over time and need to be retested whenever there is a change. This is essential for ensuring uncompromised quality. If we have modified only a small part of the system, it should be possible to reuse the existing test suite. Anyhow, for large modifications or for large systems, retesting the entire test suite will consume large amounts of time and computing resources. In this paper we propose a new method to prioritize test cases in web applications. Our test prioritization technique orders test cases in such a way that the most beneficial is executed first. Most of the existing test prioritization methods are based on the code of the system, but we propose a model-based test prioritization using activity diagram. Our technique identifies difference between original model and modified model. Using this information we plot activity paths for each test case and identify the most promising paths. The test case which covers these paths is considered as the most beneficial test cases. Our approach is effective in revealing the most promising regression test cases. We have applied our method on an online air ticket reservation system in which we could identify the most beneficial test cases from the existing ones.

Enhanced Bee Colony Algorithm for Efficient Load Balancing and Scheduling in Cloud

Cloud computing is a promising paradigm which provides resources to customers on their request with minimum cost. Cost effective scheduling and load balancing are major challenges in adopting cloud computation. Efficient load balancing methods avoids under loaded and heavy loaded conditions in datacenters. When some VMs are overloaded with several number of tasks, these tasks are migrated to the under loaded VMs of the same datacenter in order to maintain Quality of Service (QoS). This paper proposes a modification in the bee colony algorithm for efficient and effective load balancing in cloud environment. The honey bees foraging behaviour is used to balance load across virtual machines. The tasks removed from over loaded VMs are treated as honeybees and under loaded VMs are the food sources. The method also tries to minimize makespan as well as number of VM migrations. The experimental result shows that there is significant improvement in the QoS delivered to the customers.

Relative Extraction Methodology for class diagram generation using dependency graph

Requirement analysis is the preliminary step in software development process. The requirements stated by the clients are analyzed and an abstraction of it is created which is termed as requirements model. Unified Modeling Language (UML) models are helpful for understanding the problems, communicating with application experts and preparing documentation. The static design view of the system can be modeled using a UML class diagram. System requirements stated by the user are usually in natural language form despite a wide variety of formal languages and UML. This is an imprecise and inconsistent form which is difficult to be used by the developer for design. We present a new methodology for generating UML class diagrams or models from natural language problem statement or requirement specification. We have named our methodology as Relative Extraction Methodology which uses an intermediate graphical representation called dependency graph. This serves as a skeleton for the generation of UML class models. The natural language processing is done for the construction of dependency graph which is finally mapped to a class diagram.

Boundary Value Testing based on UML Models

We present a novel method to automatically generate test cases based on UML state chart specifications. In our approach, we transform the conditional predicates on state transitions and apply function minimization technique to generate the test data. We use boundary value testing methods to generate effective test cases that satisfy test coverage criteria like full predicate coverage criteria and transition path coverage criteria. Our approach achieves adequate test coverage without unduly increasing the number of test cases. The test cases are generated for class as well as cluster level testing.

Virtual Machine Placement for Improved Quality in IaaS Cloud

Cloud computing, with its immense potentials in low cost and on-demand services, is a promising computing platform for both commercial and non-commercial computation applications. It focuses on the sharing of information and computation in a large network that are quite likely to be owned by geographically disbursed different vendors. The proposed Bin packing based algorithm tries to maximize the resource utilization in a cloud datacenter and thereby increasing the profit of that cloud provider. In the proposed system, Best-fit -- Worst-fit strategy efficiently places the virtual machines to the less number of active physical servers. The jobs are scheduled using best-fit approach. The cloud broker employs worst-fit method for VM placement. The simulation result shows the significant improvements in the performance of the cloud system.

A distributed hierarchical key management scheme for mobile Ad hoc networks

In MANET, when a node is compromised it tends to reveal the other nodes key information and corrupts the whole network. The scalable method of cryptographic key management (SMOCK) proposes a method to deal with such node compromise attacks. But it has certain main drawbacks such as over dependent on centralized server and increase in key-pair when node increases. To deal with these drawbacks, we present an enhanced hierarchical key management scheme using a stable and power efficient cluster management technique. Each cluster head retains the public key of its member nodes only and act as a router when dealing with nodes of other cluster members. Using this technique, the overhead on centralized server is reduced. Moreover, the need of each node storing all public keys are diminished thus minimizing the storage overhead on each node. By Simulation results, we show that our proposed scheme achieves better delivery ratio and resilience with reduced delay and overhead.