Elastic Overlay Network Using Affiliate PoPs for Low-Latency Access
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network path selection mechanisms fail to identify optimal paths for accessing geographically dispersed applications and services, leading to suboptimal performance and high costs due to the deployment of geographically dispersed points of presence (PoPs) to ensure traffic optimization, which is particularly problematic for latency-sensitive applications.
Innovation Solution
An elastic overlay network is generated by strategically placing relays and constructing layer-two tunnels to optimize cross-domain path performance, utilizing machine learning to model path performances and dynamically select affiliate nodes as PoPs based on performance metrics, thereby creating a flexible and cost-effective network presence.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If points of presence (PoPs) are deployed at various geographic locations to support access to online applications or services, then access to geographically dispersed resources is enabled, but deployment time and resource consumption increase significantly
Solution Approach 1:
The patent creates virtual copies of network nodes and connections through overlay networking. Instead of physically deploying new PoPs, the system replicates network functionality through software-defined virtual nodes that mirror existing infrastructure, enabling rapid geographic expansion without hardware deployment
Solution Approach 2:
The patent implements nested network structures where overlay networks are built upon underlying physical networks. Virtual PoPs are nested within existing network infrastructure, allowing the system to leverage existing hardware while adding virtual layers for enhanced functionality and geographic presence
2Adaptability or versatility
If hardware is deployed to new locations to establish points of presence, then network coverage is expanded, but resource consumption and deployment cost increase
Solution Approach 1:
The system creates virtual representations of network nodes that replicate physical PoP functionality without requiring actual hardware deployment. These virtual PoPs are instantiated through software on existing infrastructure, providing network coverage expansion without proportional increases in physical hardware
Solution Approach 2:
Existing network infrastructure is made multi-functional by overlaying virtual networking capabilities. The same physical network hardware supports both traditional networking and virtual overlay networking, allowing a single infrastructure to serve multiple purposes including expanded network coverage without additional dedicated hardware
3Adaptability or versatility
If multiple network paths are established between client and geographically dispersed resources, then path performance options increase, but network complexity increases
Solution Approach 1:
The patent segments network paths into distinct overlay tunnels that can be independently managed and optimized. Each overlay tunnel represents a separate virtual path that can be selectively activated or deactivated, allowing fine-grained control over path performance without managing the complexity of the entire underlying network
Solution Approach 2:
The overlay network acts as an intermediary layer between clients and underlying physical networks. This intermediary abstraction layer simplifies path selection by presenting a simplified interface that manages multiple underlying paths automatically, shielding clients from network complexity while providing optimized routing
4Productivity
If points of presence are deployed to optimize traffic for specific locations, then traffic optimization is achieved, but the solution does not adapt to varying traffic loads over time
Solution Approach 1:
The patent implements dynamic overlay network configuration where paths and PoPs can be dynamically added, removed, or rerouted based on real-time traffic conditions. The system transitions from static network configurations to dynamic adaptive routing that responds to varying traffic loads, enabling continuous optimization without fixed infrastructure
Solution Approach 2:
The system incorporates feedback mechanisms that monitor traffic performance and load conditions, using this information to automatically adjust overlay network configuration. Performance data feeds back into path selection algorithms, enabling continuous optimization of traffic routing based on actual conditions rather than predetermined configurations
Data Source
AI summary
In one embodiment, a device obtains path probing data for paths between a set of network nodes wherein a first subset of the set of network nodes are operated by a provider of an online application or service and a second subset of the set of network nodes operated by affiliates. The device generates, based on the path probing data, a performance model of path performances between clients in different geolocations and the online application or service via each of the set of network nodes. The device determines, based on the performance model, a particular node in the second subset should be used as a point of presence for clients located in a particular one of the different geolocations to access the online application or service. The device configures a network overlay including the first subset and the particular node as points of presence for the online application or service.