Network bandwidth

Autoscaling using network bandwidth resources

The CAST AI Autoscaler enables workloads to select nodes with suitable network bandwidth specifications. For example, if a workload specifies a requirement of up to 1000 Megabits per second (Mbps) in bandwidth, the Autoscaler will pick an instance type that provides at least 1000 Mbps for inbound and at least 1000 Mbps for outbound traffic.

Supported providers

How it works

Workloads can request network bandwidth by specifying it in the container's resources stanza. The Autoscaler will respect the specified value and select a suitable instance type that guarantees sustainable bandwidth as specified by the cloud service provider.

To specify the required network bandwidth, add the scheduling.cast.ai/network-bandwidth annotation to both the requests and limits sections of the container spec. The value must be expressed in Megabits per second (Mbps).

Workload configuration examples

resources:
  requests:
    scheduling.cast.ai/network-bandwidth: "750" # Requested network bandwidth in Mbps
  limits:
    scheduling.cast.ai/network-bandwidth: "750" # Maximum network bandwidth limit in Mbps

apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    app: nginx
  name: nginx
spec:
  replicas: 1
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      labels:
        app: nginx
    spec:
      containers:
        - image: nginx
          name: nginx
          resources:
            requests:
              scheduling.cast.ai/network-bandwidth: "750"
            limits:
              scheduling.cast.ai/network-bandwidth: "750"