Real-Time Management and Control of Monitoring Elements In Dynamic Cloud Network Systems

This paper explores new scenarios where Cloud Network Service Providers take advantage of using more flexible resource management and orchestration solutions in the form of dynamic virtualised compute, network and storage resources. The main focus of this work is to analyse how those challenges will considerably impact the requirements of the monitoring process. A framework in the context of 5G is here presented to support the dynamic on-demand management, configuration and control of a monitoring subsystem which: can easily scale up / down according to the number of running entities in the system as result of the instantiation / termination of multiple services; can provide mechanisms to dynamically activate / deactivate its constituent elements on-demand according to the type of services to be monitored; and can provide mechanisms to dynamically adjust the configuration if its elements. Experimental outcomes, where a Monitoring Controller was used to adjust the measurement collection / sending rate of the probes in the monitoring subsystem on the-fly are also presented. The paper shows how this prevented the transmission of vast amounts of data when the number of virtual entities and related monitoring probes in the system scaled up to hundreds of elements.