Linda in the Fog

Défense de mémoire de Maxime Bever

The amount of data sent to data centres is constantly increasing. Additionally, the number of ‘Internet of Things’ sensors generating data is also increasing. To manage this data, the fog computing paradigm has been invented. The aim of fog computing is to create a new layer of data management before sending it to the cloud. Furthermore, and in addition to avoiding sending too much data to the cloud, the fog layer will decrease the response time of requests that could already be implemented within this new layer. For these requests to be efficient, the coordination of the message is particularly important. Accordingly, a coordination language could be the solution. In this work, the different criteria for implementing a coordination language in fog computing are explained. These criteria concern security, scalability, performance, management of low-resource devices, data placement policy and data persistency. A proof-of-concept of these criteria is presented, and its development is explained. To prove the effectiveness of this proof-of-concept, performance tests were performed, and different use cases applying the implementation in various contexts are presented.

 

Keywords: fog computing, coordination language, edge computing, Linda, Rust

Contact : Isabelle Daelman - isabelle.daelman@unamur.be
Télecharger : vCal