| Full text | |
| Author(s): |
Cardoso, Wallace
;
Martins, Eliane
;
Laranjeiro, Nuno
;
Antunes, Nuno
;
IEEE
Total Authors: 5
|
| Document type: | Journal article |
| Source: | 2019 9TH LATIN-AMERICAN SYMPOSIUM ON DEPENDABLE COMPUTING (LADC); v. N/A, p. 10-pg., 2019-01-01. |
| Abstract | |
OpenStack, the widely used cloud system nowadays, is the basis for many different user services (e.g. markets, banks, education services). Even in high available versions, OpenStack failures may potentially get its robustness down and impair user services. Faults may happen from many sources (e.g. hardware, API, messages, third-party code), and in this way, it is hard or even impossible to generate a good set of robustness tests manually (a practical approach used in the community). In this paper, we automate the robustness test generation using model-based testing and evaluate its effectiveness using the test oracle that was generated using an OpenStack's behavioral model (valid by the documentation). On the best of our knowledge, previous works do not focus on injecting faults in data structures within messages that are in transport among OpenStack services. Our results show that many of the faults we used may affect the OpenStack users, that is, these failures cause inconsistencies and misleading behavior in the system. (AU) | |
| FAPESP's process: | 17/21773-9 - Development and Testing of Resilient Systems-of-Systems: a Software Architecture Perspective |
| Grantee: | Cecília Mary Fischer Rubira |
| Support Opportunities: | Regular Research Grants |