AI Traffic Distribution Across Autonomous System Links for Load Balancing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge in autonomous systems is optimizing traffic distribution across interconnecting links to ensure balanced load and adherence to agreements, as current methods are complex, error-prone, and prone to suboptimal states due to manual or rule-based optimizations.
Innovation Solution
A method utilizing artificial intelligence to learn and optimize traffic distribution by receiving information about neighboring autonomous systems, determining a reward function result, and triggering actions to influence link load states, including modifying BGP attributes like MED, AS PATH, and PREFIX, with iterative learning to achieve optimal load balancing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual optimization actions are implemented to modify BGP attributes or prefix advertisements, then traffic distribution can be optimized, but the process becomes complex and error-prone
Solution Approach 1:
The system implements self-service through automated monitoring and optimization of traffic distribution. The autonomous system automatically detects link load states, evaluates optimization opportunities using the reward function, and executes BGP attribute modifications without manual intervention, thereby improving efficiency while reducing operational complexity
Solution Approach 2:
The system employs continuous feedback loops where traffic distribution outcomes are monitored and fed back into the optimization process. The reward function evaluates the results of previous optimization actions and uses this feedback to guide subsequent modifications to BGP attributes, creating a closed-loop control system that adapts to changing network conditions
2Extent of automation
If rule-based automation is used to select optimization actions, then manual effort is reduced, but inappropriate actions may be executed and cause oscillations between suboptimal states
Solution Approach 1:
The system dynamically changes BGP parameters (attributes and prefix advertisements) based on real-time evaluation of the reward function. Instead of following fixed rules, the system adjusts parameters adaptively by identifying actions that maximize the reward function value, thereby avoiding inappropriate actions and oscillations while maintaining high automation levels
Solution Approach 2:
The optimization system transitions from static rule-based automation to dynamic adaptive automation. The system continuously evaluates the network state, determines optimal actions based on current conditions, and executes modifications that adapt to changing traffic patterns and link load states, preventing oscillations between suboptimal states
3Ease of operation
If sequential examination of BGP preferences is used to determine traffic paths, then autonomous systems can independently select best paths, but load distribution becomes unbalanced and may not comply with agreements
Solution Approach 1:
The reward function acts as an intermediary between autonomous path selection and load distribution optimization. It translates the independently selected paths into a measurable outcome that reflects overall link utilization, allowing the system to evaluate and adjust BGP attributes to achieve better load distribution while respecting autonomous decision-making
Solution Approach 2:
The system modifies BGP parameters (such as MED attributes and prefix advertisements) to influence path selection in a way that optimizes load distribution. By dynamically adjusting these parameters based on link load states and agreement requirements, the system improves link utilization efficiency while maintaining autonomous operation
Data Source
Figure 1~2
Figure 3~4
AI summary
A method is proposed for distributing traffic over a set of links (L100, L200, L300) interconnecting a local autonomous system (AS1) with a plurality of neighboring autonomous systems (AS100, AS200, AS300) of the local autonomous system (AS1), the method comprising: /a/ receiving (101) information representative of: - a load state of the set of links (L100, L200, L300); - a type of autonomous system for each of the neighboring autonomous systems (AS100, AS200, AS300); and - prefixes relating to networks hosted by the local autonomous system (AS1) and the neighboring autonomous systems (AS100, AS200, AS300), the prefixes being advertised by an autonomous system to its neighbors; /b/ determine (102), from the information received, a result of a reward function, representative of a state of said set of links;/c/ trigger (103), based on the information received and the result of the reward function, at least one action influencing the distribution of traffic across the link set (L100, L200, L300) so as to optimize the load state of the link set (L100, L200, L300); /d/ repeat (104) steps /a/ to /c/; the at least one action being identified by an artificial intelligence configured to learn to identify, based on successive results of the reward function, actions that allow the distribution of traffic across the link set (L100, L200, L300) so as to optimize the load state of the link set (L100, L200, L300).