Affiliations: [a] Department of Computer Architecture, Universitat Politècnica de Catalunya, Barcelona, Spain | [b] Department of Computer Science, Universidad de Chile, Av. Blanco Encalada, Santiago, Chile
Correspondence:
[*]
Corresponding author: Sergio F. Ochoa, Department of Computer Science, Universidad de Chile, Av. Blanco Encalada 2120, Santiago, Chile. E-mail: sochoa@dcc.uchile.cl.
Abstract: There currently are many mobile computing devices with various properties and capabilities. These devices may need to collaborate among them to allow nomad workers to perform a common activity. Unfortunately software developers in charge of creating infrastructures or applications allowing these devices to cooperate among them, do not count with clear guidelines to design such software components; particularly when these components must work in a scenario involving heterogeneous devices. This paper presents a study that tries to understand how to address collaboration among heterogeneous mobile devices, by exploring several variables affecting the process. In particular, this study explores various strategies to borrow CPU slots from peer mobile computing devices and return the favor back later on. The study outcomes indicate there is a short list of computing and network variables affecting the collaboration capability of the mobile devices. These findings have been verified using data mining techniques. Based on these findings and the lessons learned, the article presents a simulation method of computing scenarios that can help developers to determine which computing configuration would be suitable to be used in each particular work scenario. Software designers can take advantage of this simulation method and the design guidelines reported in this paper in order to develop applications able to work appropriately in heterogeneous computing scenarios.
