TrafficManager: An Overall Functional Design

[ESWZY]

Design Concept

Design concept is now available here:

https://github.com/ESWZY/lmp/blob/traffic-mgr/doc/eBPF_Supermarket/TrafficManager/CONCEPT.md

Functional Design

Based on abstractions of Kubernetes Services and Pods, as well as the modification of network request events, this project can achieve the following functionalities through operational logic:

1. Service Resolution: It directs requests straight to backend Pods, bypassing massive iptables lookup and NAT through iptables.

2. Non-intrusive Traffic Management: It offers the ability to modify traffic across Pods and Services. Specific callers can invoke particular versions of a service, facilitating traffic migration or version rolling upgrades.
3. Metric-Based Managing: By using metric inputs, it filters and eliminates abnormal Pods, preventing requests from reaching malfunctioning Pods. If all Pods are unable to function correctly, the request is denied outright.

Development Plan

• Build the basic development framework and automatic compilation pipeline.
  • Initialize basic development framework for TrafficManager #464
• Implement kernel abstraction of Service and Pod, and design corresponding maps for storage and information transfer.
  • Implement kernel abstraction of Service and Pod, and use maps for storage and information transfer #482
  • Provide map read and write interface on the Go program side and weighted back-end pod selection #490
• Implement cluster metadata analysis and map read and write update in user mode. Consider using the Kubernetes Controller's control loop to monitor changes to the current cluster and keep the metadata in the map always up to date.
  • Retrieve k8s metadata and encapsulate automatic map read and write interface #500
• Performance optimization and development framework arrangement.
  • Verify if weighted back-end pod selection is working properly #496
  • Add performance verification through testing #507
• Investigate and develop consistent hashing capabilities to achieve fast hashing and fast Pod selection.
• Investigate and develop grayscale release function of traffic, such as canary release and blue-green release, which provides cross-Service traffic modification capabilities.
  • Support action redirecting to another service for traffic manager #519
• Implement filtering out specific abnormal nodes and Pods based on external cluster monitoring information.
  • Dynamic cluster monitoring and selection weight modification for traffic manager #524
• Performance optimization.
  • Fast weighted backend selection for traffic manager #523
• Documentation and tutorials.
  • Update overall project documentation for TrafficManager #525