21 dez 2020
10:00 Master's Defense Fully distance
Theme
Generation of Adaptation Plans Based on Quality Models for Cloud Platforms
Student
Jorge Luiz Machado da Silva
Advisor / Teacher
Cecilia Mary Fischer Rubira
Brief summary
Cloud computing has brought numerous benefits in terms of cost and scale, offering support services for the provisioning of infrastructure aimed at processing and storing data, according to the demands of applications. In this environment, trustworthiness can be defined as a customer's trust in a cloud service and its provider; however, the guarantee of this property is not trivial. First, trustworthiness is generally not made up of a single quality attribute, but rather a combination of various attributes, such as data privacy, performance, reliability, etc. Secondly, customers can see a change in the level of trustworthiness required by their application at runtime due to the degradation of the cloud service. Therefore, one of the main barriers is to guarantee the non-functional properties of trustworthiness on-the-fly. This work presents a solution that monitors, at runtime, a set of quality attributes of a given application and generates adaptation plans to ensure that an adequate amount of resources is provided by the cloud to maintain its level of trustworthiness. The solution is based on quality models for calculating metrics associated with each non-functional requirement and their combination at different types of trustworthiness levels. The main contribution of the solution is to provide an approach that deals with multiple requirements at the same time (or simultaneously) during the runtime, in order to adapt the resources of the cloud to maintain the level of trustworthiness required by the application. The solution was evaluated by two experiments considering a scenario where the level of trustworthiness of the application was composed of two attributes (data privacy and performance) and the other with three (data privacy, performance and reliability). The results showed that the approach is feasible in terms of execution of the adaptation plans at runtime to certify the level of trustworthiness required by the application.
Examination Board
Headlines:
| Cecília Mary Fischer Rubira | IC / UNICAMP |
| Edna Dias Canedo | CIC / UnB |
| Edmundo Roberto Mauro Madeira | IC / UNICAMP |
Substitutes:
| Edson Borin | IC / UNICAMP |
| Genaína Nunes Rodrigues | CIC / UnB |