Support inter-pod affinity and anti-affinity at NUMA level

[caohe]

What would you like to be added?

This issue is opened to track the development of inter-pod affinity and anti-affinity at NUMA level in Kubernetes.

Why is this needed?

Currently, Kubernetes supports inter-pod affinity and anti-affinity at the node level. However, there is a growing need for extending this support to the NUMA level.

For example, high-memory bandwidth consuming pods, like workers, can impact the performance of other pods on the same NUMA node, such as parameter servers. Allocating these pods to different NUMA nodes can mitigate such interferences.

[yanxiaoqi932]

Hello, I wanna participant it 😄

[waynepeking348]

we will have community meeting to discuss this, can we post the documents here? @pendoragon

Hello, I wanna participant it 😄

[pendoragon]

we will have community meeting to discuss this, can we post the documents here? @pendoragon

Hello, I wanna participant it 😄

sure, https://bytedance.feishu.cn/docx/VXGfdiddUoemoQx54uacwSHonPd @yanxiaoqi932 feel free to edit the agenda

[yanxiaoqi932]

we will have community meeting to discuss this, can we post the documents here? @pendoragon

Hello, I wanna participant it 😄

sure, https://bytedance.feishu.cn/docx/VXGfdiddUoemoQx54uacwSHonPd @yanxiaoqi932 feel free to edit the agenda

Thank you. I have applied for permission to the online document.

[pendoragon]

v0.4.0 tbd