Zainal Hisham Che Soh

Distributed t-way test suite data generation using exhaustive search method with map and reduce framework

In this paper, a new distributed algorithm for generating test suite for t-way testing utilizing a shared tuple space server as transport layer between master and worker is presented. The design of the algorithm is based on the Map and Reduce software framework implemented using tuple space technology. The proposed algorithm requires listing all the possible interaction elements into all partitioned space to ensure maximum interaction coverage in all partition. The exhaustive test cases is generated and loaded into their respective partition using hash based routing algorithm. The final test suite is produced from a reduced the test case obtained from the worker calculation on maximum interaction coverage. An encouraging simulation result is obtained on effectiveness of using the tuple space technology to parallelize the test suite generation on single machine and multi machine for different input parameter.

Evaluation of Virtual Desktop Infrastructure host using User Workload Simulator

Increasing cost and complexity of maintaining physical computers have caused large organization to consider alternative solution such as implementing Virtual Desktop Infrastructure (VDI) in their enterprise network. However, before deploying VDI solution in a large scale, it is essential to evaluate the sufficiency of the solution first. As a result, User Workload Simulator was developed. The simulator was used to evaluate two VDI solutions that are named VDI A and VDI B. The objective of the evaluation is to compare which solution can be a better VDI server host. From the result, VDI A is sufficiently better than VDI B for VM density per host. But in terms of network bandwidth utilization and Input/Output Operations Per Second (IOPS) consumption per host, VDI B is sufficiently better as it utilized less network bandwidth and consumed less IOPS on most of the user profiles evaluated.

Distributed t-way test suite execution with code coverage on multiple machines

This paper present a distributed test suite execution in term of code coverage analysis for testing a t-way test suite data using tuple space technology with Map and Reduce mechanism. The generated t-way test suite data and software under test (SUT) source code are preloaded into their respective partition using hash based routing algorithm. Each partition parsed each test case within the test suite into actual test data file. The code coverage of SUT source code is tested with each test case in actual test data and produces a test coverage result at their respective partition in term of class, method, block and line coverage. A case study of CGPA calculator as SUT is done to measure the test suite execution performance in term of test coverage with varying interaction strength on single and multiple machine environments. The scalability analysis of distributed test suite execution is also done in term of speedup gained while running on a multiple machine environments. An encouraging result is obtained on test coverage and speedup for multiple machine environments as compared to single machine environment. Higher test coverage and more speedup is obtained in higher machine environments.