System architecture of Intelligent Monitoring in multi-domain orchestration

Orchestrating a service or resources across multiple administrative domains requires three main high level steps of operation: i) capability detection of other domains ii) placement of the service request across the domains iii) assurance that the service is functioning within the acceptable bounds of the service level agreement (SLA). This paper focuses on the assurance step; in particular, we present a novel architecture and preliminary implementation that supports monitoring of KPIs across multiple administrative domains. The challenges towards realising such an architecture are: i) coordinated monitoring with no direct access to the other domain's infrastructure ii) monitoring over an abstracted (instead of actual) topology that each administrative domain may expose to other domains iii) different domains have different systems for monitoring with different KPIs as well as different ways of measuring those KPI. Our architecture, referred to as IMoS addresses these challenges to provide an end-to-end monitoring as a fundamental functionality for supporting assurance and SLA management for services orchestration across multi-administrative-domains. In addition, the preliminary results are provided from the first proof of concept implementation of an IMoS. The work done in this paper has been developed within the H2020 ICT14 project 5G Exchange.