controller.adoc

Alcor Regional Controller Design

Table of Contents

• High-Level Architecture
  • Design Principles
• Micro-Service Framework
  • Micro-service Snapshot
  • Concurrency and Event Ordering
• Availability Zone Resilience
• Service-to-Service Communication

Note	This document is under development

High-Level Architecture

Design Principles

• Regional Scope, AZ resilience

• Simple network resource abstraction

• Loosely coupled components for flexible partitioning and easy scale out

• Top-down configuration driving towards eventual consistency

• Decoupling among services

  • Database access only through service

  • Isolation of database access and cache on the service level

  • Enable flexible partitioning for various services

Micro-Service Framework

1. One controller instance is one Kubernetes application

2. One microservice is one Kubernetes service

3. One microservice could consist of multiple service instances (stateless or stateful) to improve availability, scalability and performance

Micro-service Snapshot

Category	Micro-service Name	Acronym	Description	Type
Resource Management Services	VPC Manager	VPM	VPC lifecycle management	Stateless
	Subnet Manager	SNM	Subnet lifecycle management	Stateless
	Port Manager	PM	Port lifecycle management	Stateless
	Route Manager	RM	Route table and rule management	Stateless
	Private IP Manager	PIM	VPC private IP lifecycle management (IPv4/6)	Stateless
	Virtual Mac Manager	VMM	Virtual MAC pool management	Stateless
	Security Group Manager	SGM	Security group management	Stateless
	Virtual IP Manager	VIM	Public virtual IP management	Stateless
	DNS Manager	DM	DNS/DHCP record management	Stateless
	Node Metadata Manager	NMM	Datacenter metadata management for physical nodes/machines	Stateless
Database and Cache Services	Apache Ignite		Database services to store resource states	Stateful
Infrastructure Services	Data-Plane Manager	DPM	Responsible of sending network configuration to nodes	Stateless
	Network Health Monitor	NHM	Physical nodes/machines management for control plane, including in/out of services, health status maint	Stateless
	Gateway Manager	GM	Responsible of managing gateway	Stateless
	Resource Pre-provisioning Manager	RPM	TBD	Stateless
Messaging Services	API Gateway	GW	Responsible of request routing, composition, and protocol translation	Stateless
	Message Queue (TBD)	MQ	Messaging services for controller and agent communication	Stateful/Stateless

Concurrency and Event Ordering

Four types of concurrent network resource update:

Concurrent Event Types	Example	Approach
Operation on decoupled resources	CURD of resources under two different/unpeered VPCs	Free to update simultaneously
Operation on loosely relevant resources	Add one port, and delete the other in the same subnet	No conflict on resource management Network conf programming: Network conf versioning + version-awareness at ACA
Operation on directly coupled resources	Delete a VPC and create a subnet for an empty VPC	Timestamp issued by API gateway Check associated resource status DB cleanup for unstaged transactions
Operation on the same resource	Update operation and delete operation on the same port	Customer experience: may have different experience if executed in different order Resource management: no conflict (using DB concurrency + timestamp versioning) Network configuration programming: no conflict

Availability Zone Resilience

TBD

Service-to-Service Communication

TBD