OpenYurt v1.6 Release: Node-Level Traffic Multiplexing and Enhanced Edge Autonomy

On January 8 (Beijing time) the OpenYurt v1.6 version was released. OpenYurt, an open‑source project in the edge cloud‑native space, aims to solve the challenges of distributed compute resources and business management. It adopts a cloud‑edge‑node integrated architecture built on Kubernetes in a non‑intrusive way, offering core capabilities such as edge autonomy, cross‑region communication, multi‑region resource and application management, and device management. The v1.6 release mainly adds node‑level traffic multiplexing and enhanced edge autonomy.

In an OpenYurt cluster, control‑plane components are deployed in the cloud while edge nodes usually interact with the cloud control plane over the public network. List/watch resources (e.g., Kubelet, Flannel, Kube‑proxy, CoreDNS) require a replica on each edge node, which creates significant cloud‑edge traffic as the cluster scales. To alleviate this, YurtHub provides a traffic‑multiplexing module that proactively fetches specific resources from the kube‑apiserver and caches them. When a client requests these resources through YurtHub, the data is served from the cache instead of being proxied to the kube‑apiserver, reducing cloud‑edge traffic by roughly 50% in large‑scale Pods and Services deployment scenarios.

The request flow works as follows:

MultiPlexer Cache requests list/watch from the kube‑apiserver for the targeted resources and stores them in memory.

The client requests the YurtHub server for the resources.

YurtHub determines whether the resource is eligible for multiplexing.

If it is, YurtHub returns the data from the local MultiPlexer Cache, reducing cloud‑edge traffic; otherwise it forwards the request to the kube‑apiserver.

OpenYurt already provides strong edge autonomy, ensuring that applications on edge nodes continue to run even when the cloud‑edge network is disconnected. However, the existing autonomy has limitations: pods are never evicted when a node uses the autonomy annotation, even if the node fails, and the autonomy feature requires disabling the NodeLifeCycle controller in the Kube‑Controller‑Manager, which is not feasible in managed Kubernetes environments.

Version 1.6 enhances autonomy with two main improvements:

A new annotation node.openyurt.io/autonomy-duration lets users specify how long a node may remain autonomous. If the node stops sending heartbeats for less than this duration, the system treats it as a network issue and does not evict pods; if the duration is exceeded, the system assumes a node failure and evicts the pods.

The NodeLifeCycle controller is no longer disabled in the Kube‑Controller‑Manager. Instead, Yurt‑Manager adds an Endpoints/EndpointSlices webhook that keeps pod endpoints marked as Ready during autonomy, preventing them from being removed from Services. The webhook workflow is illustrated in the diagram below.

Other updates include links to the GitHub release page for detailed change logs and author contributions.

The new capabilities of OpenYurt v1.6 are already available in the ACK Edge product, and the community is actively developing v1.7, which will focus on node‑pool traffic multiplexing and deploying Kubernetes clusters locally with OpenYurt. Interested users are invited to join the community via GitHub, DingTalk (group ID 12640034121), the weekly community meeting, or the Slack channel.

Related links:

GitHub release: https://github.com/openyurtio/openyurt/releases

ACK Edge: https://help.aliyun.com/zh/ack/ack-edge/product-overview/ack-edge-overview

GitHub issues: https://github.com/openyurtio/openyurt/issues

Community weekly meeting: https://meeting.dingtalk.com/

Slack channel: https://openyurt.slack.com/unsupported-geo