Configuration/Infrastructure-aware testing of MapReduce programs
The implemented programs in the MapReduce processing model are focused in the analysis of large volume of data in a distributed and parallel architecture. This architecture is automatically managed by the framework, so the developer could be focused in the program functionality regardless of infrast...
| Autores: | , , , , , |
|---|---|
| Tipo de recurso: | artículo |
| Fecha de publicación: | 2017 |
| País: | España |
| Institución: | Universidad de Oviedo (UNIOVI) |
| Repositorio: | RUO. Repositorio Institucional de la Universidad de Oviedo |
| Idioma: | inglés |
| OAI Identifier: | oai:digibuo.uniovi.es:10651/40737 |
| Acceso en línea: | http://hdl.handle.net/10651/40737 |
| Access Level: | acceso abierto |
| Palabra clave: | Software testing Functional testing Big Data Engineering Hadoop |
| Sumario: | The implemented programs in the MapReduce processing model are focused in the analysis of large volume of data in a distributed and parallel architecture. This architecture is automatically managed by the framework, so the developer could be focused in the program functionality regardless of infrastructure failures or resource allocation. However, the infrastructure state can cause different parallel executions and some could mask the faults but others could derive in program failures that are difficult to reveal. During the testing phase the infrastructure is usually not considered because commonly the test cases contain few data, so it is not necessary to deploy a parallel execution or handle infrastructure failures, among others potential issues. This paper proposes a testing technique to generate and execute different infrastructure configurations given the test input data and the program under test. The testing technique is automatized by a test engine and is applied to real world case studies. As a result, the test engine generates and executes several infrastructure configurations, revealing a functional fault in two programs |
|---|