RESOURCE AVAILABILITY TRANSPARENCY Clause Samples

RESOURCE AVAILABILITY TRANSPARENCY. With the current technology, each microservice can use only the resources that are inside its own cluster. This statement holds for both normal operations (e.g., when the service is started), and when an update (e.g., automatic scaling) is requested. This behaviour prevents a service from using available resources located in other parts of the continuum, hence possibly ending up in a service disruption even in the presence of available resources, but located elsewhere in the continuum. This problem is not very important in cloud datacenters, where it is very unlikely to run into a lack of resources. Even in the case of small clusters obtained by acquiring only a small subset of the datacenter resources, the problem is irrelevant thanks to the capability to (automatically) resize the cluster, e.g., by adding/removing worker nodes to the given cluster. Instead, this problem is more important when considering a small pool of resources available at the edge, in which usually a few servers are available. The capability to acquire new physical resources, when needed, can only be achieved by leveraging other worker nodes possibly available in the closest vicinity. This section presents the main interactions envisioned in FLUIDOS and the main components in the FLUIDOS architecture.