Application Controlled Path Selection Over Transit Providers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network routing technologies, such as BGP, lack control over path selection at the application or server level, leading to unbalanced utilization of paths, potential performance issues due to lack of performance metric advertisement, and slow response to failures, which can result in suboptimal content delivery.
Innovation Solution
Implementing application-controlled path selection by advertising unique address prefixes for each transit provider path, allowing servers to proactively select paths based on configurable policies and performance metrics, thereby overriding standard BGP operations without disabling router intercommunication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If BGP path selection is used for routing traffic, then path selection is automated at the network level, but servers have no control over which path their traffic uses and path utilization becomes unbalanced
Solution Approach 1:
The patent segments path selection control by introducing application-level path selection that operates independently from network-level BGP routing. Each server can select specific paths for its traffic while BGP continues to operate autonomously at the network level, dividing the control function between two layers rather than having a single centralized control mechanism
Solution Approach 2:
The patent adds a new dimension of control by implementing path selection at the application layer (server level) in addition to the existing network layer (BGP level) path selection. This creates a multi-dimensional control architecture where both levels operate simultaneously, allowing servers to influence path selection without disabling BGP automation
2Device complexity
If BGP routers select paths based on hop count, then routing is simplified, but performance metrics such as latency and packet loss are not considered
Solution Approach 1:
The patent implements preliminary path performance assessment by having servers evaluate multiple available paths before selecting one for traffic transmission. Servers can proactively measure or estimate performance metrics (latency, packet loss, throughput) and choose the optimal path in advance, rather than relying solely on BGP's hop-count-based selection that does not consider actual performance
Solution Approach 2:
The patent enables servers to autonomously perform path selection based on their own performance requirements and observations. Each server independently evaluates path performance and makes routing decisions for its traffic, eliminating the need for complex BGP router configurations while improving reliability through application-aware path selection
3Productivity
If BGP routers rely on neighboring router advertisements, then routing information is exchanged efficiently, but response to path failures is slow
Solution Approach 1:
The patent implements feedback mechanisms where servers continuously monitor path performance and can detect failures directly at the application level. When a server detects path degradation or failure through performance metric monitoring, it can immediately switch to an alternative path without waiting for BGP failure propagation, creating a closed-loop feedback system that responds faster than traditional BGP routing
Data Source
AI summary
Some embodiments move the task of selecting between different transit provider paths from the network level to the application level. Some embodiments perform network level configurations involving a destination network router advertising over a first transit provider path, a unique first address identifying a destination network server as reachable via the first path and advertising over a second transit provider path, a unique second address identifying the destination network server as reachable via the second path. Some embodiments further perform application level configurations involving a source network server passing a first packet to the destination network server over the first path by addressing the first packet to the first address and passing a second packet to the destination network server over the second path by addressing the second packet to the second address. The path selection may be based on policies accounting for congestion, performance, and other metrics.