Redundant Route Controllers in Global Network Access Points

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing load balancing techniques face challenges such as creating single points of failure, delays in propagating changes, and loss of control over request distribution, particularly in DNS and anycast configurations.

Innovation Solution

A distributed set of global access points using anycast routing techniques to advertise global network addresses, selecting endpoints based on load balancing criteria, and employing NAT and encapsulation to route requests efficiently and resiliently across multiple geographic locations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If a load balancer is used to distribute requests among servers, then resource usage is balanced, but a single point of failure is created

Engineering Contradiction:
Improveresource usage balanceVSAvoidsingle point of failure
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The system segments the load balancing function across multiple distributed access points instead of using a single centralized load balancer. Each access point independently performs load balancing for its assigned subset of clients, eliminating the single point of failure while maintaining resource distribution capabilities

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces anycast addresses as an intermediary mechanism that enables clients to reach multiple access points simultaneously. The anycast routing infrastructure acts as a mediator that distributes client connections across different access points, preventing any single access point from becoming a critical failure point

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If DNS load balancing is used to distribute requests, then redundancy is improved, but delay in propagating changes occurs

Engineering Contradiction:
ImproveredundancyVSAvoidpropagation delay
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system pre-establishes anycast addresses and routing configurations across multiple access points before load balancing is needed. This preliminary setup ensures that when load distribution changes are required, the infrastructure is already in place to rapidly redirect traffic without waiting for DNS propagation

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent replaces the DNS-based load balancing mechanism with a direct anycast routing mechanism. Instead of relying on DNS resolution and caching systems that introduce propagation delays, the system uses network layer anycast routing that provides immediate load distribution changes

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Ease of operation

If anycast routing is used to distribute requests, then control over request distribution is improved, but volatility in load balancing occurs due to neighboring network changes

Engineering Contradiction:
Improvecontrol over request distributionVSAvoidload balancing stability
Core Design Contradiction:
Ease of operationVSStability of the object's composition

Solution Approach 1:

The system dynamically adjusts routing parameters and load balancing weights based on real-time monitoring of access point status and network conditions. This allows the system to maintain stable load distribution even when neighboring network configurations change, by adapting to the new conditions rather than being disrupted by them

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentUS10855580B2Consistent route announcements among redundant controllers in global network access point
Publication Date: 2020.12.01 AMAZON TECH INC
  • US10855580B2 patent drawing
  • US10855580B2 patent drawing
  • US10855580B2 patent drawing

AI summary

Systems and methods are described to enable management of redundant route announcements in an access point including multiple packet processors. Route controllers are described that can generate routing information distributing incoming packets to the access point among the packet processors. The route controllers can operate redundantly, such that a failure of a single controller does not cause a complete failure of the access point. To avoid different announcements by different route controllers (particularly under partial failure scenarios), the route controllers utilize a strongly consistent data store to store routing information. So long as a record within the data store contains valid information, it is considered authoritative and routing information from the record is announced by all route controllers. If the information is invalid, the route controllers attempt to overwrite the information with new, higher priority routing information. A single write succeeds, resulting in new authoritative routing information.