Dynamic Link Weight Path Computation for Network Congestion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing path computation methods in on-demand networks fail to effectively manage congestion and throughput, as they do not consider current and future network status, leading to potential link blocking and throughput collapse, especially during peak times or when future bandwidth requirements are not accounted for.
Innovation Solution
A method that determines dynamic link weights based on current and future network status, assigning high values to links with insufficient capacity, and uses these weights to select paths that minimize cost and prevent network-wide congestion, incorporating future capacity plans and traffic forecasts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If adaptive path computation is used to avoid congested links by adding penalties, then link blocking is reduced, but throughput collapse occurs when too many connections are routed on long paths
Solution Approach 1:
The patent implements dynamic link weights that automatically adjust based on real-time network status and future forecasts. During congestion, the system dynamically modifies path selection criteria to balance load distribution, preventing both link blocking and throughput collapse by adapting to changing network conditions rather than using static penalty-based approaches
Solution Approach 2:
The system incorporates feedback mechanisms that monitor current network status and future traffic forecasts, using this information to adjust path computation decisions. This closed-loop approach allows the network to respond to congestion patterns and prevent throughput collapse by learning from past performance and anticipating future conditions
2Speed
If shortest-path path computation is used to maximize efficiency, then connection setup time is reduced, but bottleneck links become overloaded causing blocking during congestion
Solution Approach 1:
The system transitions from static shortest-path computation to dynamic path selection that adjusts in real-time based on network status. The dynamic link weights maintain computational efficiency while preventing bottleneck overload by adapting path preferences according to current and forecasted network conditions
Solution Approach 2:
The patent changes the parameters used in path computation from fixed metric-based weights to dynamic weights that incorporate network status and future forecasts. This allows the system to maintain shortest-path efficiency when appropriate while avoiding congestion by modifying path selection parameters during high-load conditions
3Ease of operation
If connections are allocated on bottleneck links to satisfy immediate demands, then short-term connection requirements are met, but long-term connections are blocked and throughput is reduced
Solution Approach 1:
The system performs preliminary actions by forecasting future network status and pre-adjusting link weights before congestion occurs. This anticipatory approach allows the network to prepare for future demands and allocate resources in advance, preventing both short-term blocking and long-term throughput reduction
Solution Approach 2:
The patent dynamically changes link weight parameters based on connection duration and network status. Short-term connections may be routed through bottleneck links when appropriate, while long-term connections are guided toward alternative paths, optimizing resource allocation for different connection types and time horizons
Data Source
AI summary
A method for path computation in a network includes determining dynamic link weights for links in the network, responsive to a request for a path, for a connection, between a source node and a destination node in the network with a requested bandwidth amount, wherein the dynamic link weights, for each link, are based on a current status of the link and a future status of the link; determining one or more paths for the request based on the dynamic link weights; and selecting a path of the one or more paths to minimize cost in the network. The method can be implemented through a Software Defined Networking (SDN) controller.