Edge Compute Placement via Latency-Based Selection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Traditional cloud computing deployments in data centers face limitations in achieving low latency due to network distance and number of hops between end-user devices and computing resources, hindering responsive applications like game streaming, virtual reality, and autonomous vehicles.

Innovation Solution

Deploying cloud provider network compute resources within communications service provider networks, known as edge locations, to reduce latency by bringing computing resources closer to end users, utilizing provider substrate extensions that integrate with CSP networks for low-latency access.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If cloud computing resources are deployed in centralized data centers, then resource pooling and management efficiency are improved, but network latency increases due to distance and number of hops between end-user devices and computing resources

Engineering Contradiction:
Improveresource management efficiencyVSAvoidnetwork latency
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent segments the centralized cloud computing infrastructure into distributed edge computing nodes deployed within communications service provider networks. These edge locations are strategically positioned closer to end users, dividing the monolithic data center architecture into multiple geographically dispersed computing resources that reduce network latency while maintaining centralized management capabilities through the virtual machine migration system.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a new spatial dimension to cloud deployment by placing computing resources within communications service provider networks rather than traditional data centers. This dimensional shift from centralized geographic locations to distributed network-integrated locations enables low-latency access while preserving the economic benefits of resource pooling through virtualization and migration capabilities.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Ease of operation

If compute resources are preselected based on current client assignments, then resource allocation simplicity is improved, but latency optimization is worsened because the only selection flexibility is gateway choice, not compute node selection

Engineering Contradiction:
Improveresource allocation simplicityVSAvoidaccess latency
Core Design Contradiction:
Ease of operationVSLoss of time

Solution Approach 1:

The patent implements dynamic compute resource selection through virtual machine migration capabilities that allow compute instances to be moved between different edge locations based on real-time latency requirements. This dynamic approach replaces static preassignment with flexible, on-demand resource allocation that optimizes latency while maintaining operational simplicity through automated migration management.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system incorporates feedback mechanisms that monitor latency performance and trigger virtual machine migrations when latency thresholds are exceeded. This feedback-driven approach enables automated optimization of compute resource placement based on actual network conditions, balancing operational simplicity with latency optimization through continuous monitoring and adaptive resource relocation.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP4049139B1Latency-based placement of cloud compute instances within communications service provider networks
Publication Date: 2024.04.17 AMAZON TECH INC
  • EP4049139B1 patent drawingFigure 1
  • EP4049139B1 patent drawingFigure 2
  • EP4049139B1 patent drawingFigure 3

AI summary

Techniques for launching compute instances on cloud provider network substrate extensions deployed within communications service provider networks are described. A service of a cloud provider network receives a request to launch a compute instance from a customer, the request including a latency requirement. A provider substrate extension is selected to host the compute instance from a plurality of provider substrate extensions of the cloud provider network based at least in part on the latency requirement. The plurality of plurality of provider substrate extensions are connected to a communications service provider network and controlled at least in part by the service of the cloud provider network via a connection through the communications service provider network. A message is sent to cause the selected provider substrate extension to launch the compute instance for the customer.