Home // International Journal On Advances in Internet Technology, volume 15, numbers 3 and 4, 2022 // View article

New Software Architecture for Monitoring Mobile Applications

Authors:
Fabrice Mourlin
Guy Lahlou Djiken
Laurent Nel

Keywords: Big Data; indexing; log analysis; distributed application; AI model; storage efficiency; anomaly detection, explanatory report.

Abstract:
The monitoring activity remains an activity that disturbs the system under control. We all try to minimize these disturbances in order to observe a behavior as close as possible to reality. In IT, this requires the implementation of a specific software architecture. Our use case concerns the monitoring of embedded applications on mobile devices for which the collected data sometimes contain errors that we want to explain. To this end, we seek to trace the important events of our calculations in order to qualify the anomalies in our processing. We have implemented a monitoring layer within mobile applications in order to perform intelligent monitoring on a set of mobile devices. We have defined a Big Data workflow to collect, index and store log data for submission to an artificial intelligence (AI) model. A crucial aspect of this collection relies on the use of partitioned topics and thus a better distribution of the data. With the increase of the data flow to be processed, the performance remains insufficient and we have opted for a persistence layer adapted to our data processing. We detect behavioral anomalies through the analysis of software logs deployed on embedded devices. Based on the patterns recognized in the logs, our AI model provides us with a sequence of system operations. These operations are then scheduled to redeploy a service, change a driver, perform a library update, etc. In the end, we build management reports every week for the maintenance team. These documents help track maintenance activities. They provide a record of important events such as equipment downtime or removal of obsolete services.

Pages: 32 to 42

Copyright: Copyright (c) to authors, 2022. Used with permission.

Publication date: December 31, 2022

Published in: journal

ISSN: 1942-2652