Analysis and Remediation of Interoperability problems of an AMQP based system
Abstract
To achieve a high level of communication at different layers independently of a specific actor, interoperability represents the current interest for all types of organizations for its open standard aspect capable of performing collaboration on different concerns. Indeed, interoperability allows the exchange of information along several axes and the use of this information to achieve common objectives. Dysfunctions related to these exchanges can have a negative impact on the overall performance of communicating. Therefore, interoperability should be continuously verified and improved. We present in this paper a monitoring considering interoperability at the data level, through the information collect about existing data exchanges that use the messaging based broker RabbitMQ as a mean of communication. The information collected supplies a set of indicators that are confronted with interoperability requirements allowing us to detect interoperability problems and provide their potential causes. Indicators characteristics offer various exploitations depending on their nature and utility.We use AHP to provide an analysis of the importance of these indicators and give an ordering of the potential causes of interoperability problems based on experts' judgments. We implemented our framework on top of the data analysis platform: Moose and experimented it on a Publish-Subscribe system implemented with the broker RabbitMQ using the protocol AMQP.
Origin : Files produced by the author(s)