SWM QoS Scheduler
The SWM QoS Scheduler is an extension of the Kubernetes scheduler that
implements an extended model for describing enhanced application requirements
and infrastructure capabilities, according to the Seamless Computing concept.
“QoS” is short for “quality of service”.
The SWM QoS Scheduler considers the extended application (QoS) requirements
when placing application components (further called workloads) to compute
nodes in the Kubernetes cluster. The main enhancements beyond the existing
Kubernetes scheduler are:
- schedule multiple workloads (Kubernetes pods) at once, considering
dependencies between these workloads
- network-aware scheduling: make Kubernetes aware of the network topology,
connectivity, and resource availability between worker nodes, and consider
this in the scheduling decision (according to the communication requirements
of the applications)
Getting Started
For installing the SWM QoS Scheduler (including all CRDs and required
containers) and to deploy a small demo application, see quick start.
License
All code files are licensed under Apache license version 2.0. All
documentation is licensed under Creative Commons Attribution-ShareAlike 4.0
International.