Execution pipeline

Contexts in source publication

Context 1

... tags like a location tag "Germany" can be assigned to a node. This tag- ging can be defined in a cluster definition file and will be assigned to a node in the install node step shown in Figure 4. Docker Swarm would schedule a CouchDB database container only on nodes which are tagged as "location=Germany" if a constraint filter is applied like shown. ...

Context 2

... Section presents a software prototype that is implemented in Ruby and provides a command line tool as core component. The tool can be triggered in the execution phase of a MAPE auto-scaling loop [9] and scales elastic container platforms according to an execution pipeline (see Figure 4). The control process interprets cluster description format (the intended state of the container cluster, see Appendix A) and the current state of the cluster (attached nodes, existing security groups, see Appendix B). ...

Context 3

... control theory inspired execution control loop compares the intended state of an elastic container platform with the current state and derives necessary actions. These actions are processed by the execution pipeline explained in Figure 4. The elastic con- tainer platform is operated in a set of synchronized security groups across di↵erent IaaS infrastructures. ...

Context 4

... proposed prototype can work with arbi- trary roles and rolenames. Role-specifics can be considered by Platform drivers (see Figure 3) in their install, join and leave cluster hooks (see Figure 4). A typical Deployment can be expressed using this JSON snippet. ...

Context 5

... execution pipeline of the control loop derives a prioritized action plan to reach the intended state (see Figure 4). The reader should be aware that the pipeline must keep the a↵ected cluster in a valid and operational state at all times. ...