Does Kubernetes use Consistent Hashing?

Yes, Kubernetes uses consistent hashing for load balancing and distributing traffic across pods in a Kubernetes cluster. When a client sends a request to a service, the request is directed to one of the pods behind the service.

• Kubernetes uses consistent hashing to map each request to a specific pod, ensuring that requests for the same resource are consistently routed to the same pod.
• This helps in maintaining session affinity and ensures that requests related to the same session are processed by the same pod.

Below is an explanation of how it works:

1. Service Definition

When you create a service in Kubernetes, you define a set of pods (backends) that the service should route traffic to. Each pod has its own IP address and port.

2. Service IP and Port

Kubernetes assigns a cluster-internal IP address and port to the service. Clients within the cluster can use this IP and port to access the service.

3. Hashing

When a client sends a request to the service, Kubernetes calculates a hash of the request’s attributes (such as source IP address, destination IP address, source port, destination port, protocol, etc.).

4. Pod Selection

Kubernetes uses this hash to select a backend pod for the request. The hash value determines which pod will receive the request. This is where consistent hashing comes into play.

5. Consistent Hashing

Consistent hashing ensures that even if the number of pods changes (pods are added or removed), the majority of requests are still routed to the same pods. This is achieved by mapping each pod to a point on a hash ring (a virtual ring). The hash of the request is used to find the nearest point on the ring, and the corresponding pod is selected.

6. Session Affinity

Consistent hashing also helps in maintaining session affinity (also known as sticky sessions) where requests from the same client are consistently routed to the same pod. This is useful for applications that require stateful communication with clients.

Overall, Consistent Hashing in Kubernetes ensures that requests are distributed evenly among pods while maintaining session affinity and resilience to pod changes.