A Topologically Aware Resource Management System

Abstract: As companies fight for market share whoever is able to bring products to market faster has an advantage over their competitors. Therefore it is absolutely essential to constantly evaluate and optimize processes to achieve shorter time-to-market for products. These optimizations have to be carried out in all parts of a company. This thesis describes one such attempt made by a Swedish telecommunication vendor focused on enabling a resource management system to gain a greater understanding of the resources available during testing. This system manages all of the hardware utilized by the users, software testers, within one particular part of the organization and aids users by automatically converting the information stored in its database into a configuration file that will later be used in the testing framework’s execution environment. Unfortunately, the current version of this resource management system lacks semantic understanding of the information necessary to automatically generate the configuration file, leaving a rather large part of the configuration file to be manually entered by the testers, a rather time-consuming task. The inability to completely automate the process means that the testing process is slower, more error prone, and increases the work needed for a new engineer to become a productive software tester. In order for the resource management system to automatically generate the configuration file it needs to know not only which resources it is managing, but must also how these resources are interconnected, i.e. the topology of the resources. For this reason this thesis describes how to make the resource management system topologically aware, thus making verification of the System under Test (SUT) more efficient and mitigating the problems mentioned above. This thesis does not deal with the intricate details of how to automatically extract the topology, as this is inherently domain specific and thus difficult to generalize. Rather, this thesis focused on how to allow users to custom-build their desired topology by defining a set of rules that restrict how resources can be interconnected. The goal of providing functionality for storing and retrieving topological information from database has been successfully achieved, and the resulting code has been integrated into the existing resource management system. However, the functionality has not yet been delivered because of a limitation in our front controller that stops us from providing an efficient web interface to our tool. After delivery the implemented solution is expected to remove most manual work related to test configuration and therefore also reduce the learning curve for new engineers.