Following previous articles posted on this blog about new operational models, microservices orchestration in Edge and Cloud, and SLAs use, this blog entry presents one of the main features implemented in Pledger during the last year: the offloading of microservices at runtime from Edge to Cloud or from Cloud to Edge based on SLAs and a continuous QoS monitoring.
The idea is to migrate all or part of the microservices deployed in an infrastructure (e.g., Edge device with Docker) to a more suitable infrastructure (e.g., Cloud with Kubernetes) when the agreed SLA is not met. For example, if an Edge device where several microservices are running is being stressed (e.g., high CPU or Memory usage), and the correspondent SLA is being violated (the monitored metrics don't meet the defined QoS for this SLA), Pledger can take the decision of moving at runtime all or part of these microservices to an infrastructure where the SLA can be met.
The same way Pledger takes these migration / offloading decisions at runtime based on SLAs and QoS violations, it can also take other kind of decisions, like to increase or decrease the horizontal and vertical scalability of the different microservices individually.
All these features will help applications owners and service providers to meet the expected QoS, while also benefiting application consumers by improving the QoE.