Congestion-Based Notification for Stateful PCE Fast Reroute
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Stateful Path Computation Elements (PCEs) face challenges such as high signaling churn and scalability issues due to the need for frequent path recomputation and state maintenance during network failures, leading to inefficient rerouting and potential network instability.
Innovation Solution
Implement a congestion-based notification mechanism that dynamically determines the level of congestion in backup tunnels after a network element failure, triggering rerouting requests only when congestion exceeds a threshold, thereby reducing unnecessary path computation requests to the PCE.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If stateful PCE maintains tunnel states and performs frequent path recomputation during network failures, then path computation accuracy is improved, but signaling churn increases and scalability deteriorates
Solution Approach 1:
The patent implements preliminary action by pre-computing backup paths and maintaining tunnel state information in the stateful PCE before failures occur. When a failure is detected, the pre-computed backup paths can be quickly activated without requiring extensive real-time computation, thus maintaining path computation accuracy while reducing signaling churn during actual failure events
Solution Approach 2:
The patent implements feedback mechanisms where the stateful PCE receives tunnel state reports from PCCs and uses this feedback to make intelligent decisions about when to trigger path recomputation. The PCE monitors congestion levels and failure conditions, and only initiates path recomputation when necessary, thereby maintaining accuracy while improving scalability by avoiding unnecessary signaling
2Reliability
If the PCE is triggered to recompute paths for every tunnel using a backup tunnel, then network reliability is improved, but signaling churn increases and processing delays worsen
Solution Approach 1:
The patent applies local quality by implementing selective path recomputation based on local conditions. Instead of triggering recomputation for all tunnels uniformly, the system evaluates local congestion levels and failure conditions at each PCC and only triggers recomputation for tunnels where it is actually needed. This maintains network reliability for affected tunnels while avoiding unnecessary processing delays for others
Solution Approach 2:
The patent implements partial action by performing path recomputation only for the subset of tunnels that require it, rather than for all tunnels. The stateful PCE identifies which specific tunnels are affected by failures or congestion and triggers recomputation only for those, thereby maintaining reliability for critical tunnels while significantly reducing overall processing delays and signaling churn
3Productivity
If congestion notification is sent to all head-ends, then network optimization is improved, but signaling overhead increases and scalability deteriorates
Solution Approach 1:
The patent implements segmentation by dividing the network into segments managed by individual PCCs. Each PCC independently monitors congestion conditions for its local tunnels and only triggers path recomputation requests to the PCE when congestion is detected locally. This segmented approach maintains network optimization by ensuring congestion is addressed while reducing signaling overhead by avoiding unnecessary notifications to all head-ends
Solution Approach 2:
The patent applies self-service by enabling PCCs to autonomously monitor their own tunnel states and make local decisions about when to request path recomputation. Each PCC serves itself by maintaining local tunnel state information and independently determining when congestion notification is necessary, thereby improving network optimization while reducing the signaling overhead that would result from centralized monitoring of all tunnels
Data Source
AI summary
In one embodiment, once activation of use of a backup tunnel is detected for a primary tunnel, then a level of congestion of the path of the backup tunnel may be determined. In response to the level being greater than a threshold, a head-end node of the primary tunnel is triggered to reroute the primary tunnel (e.g., requesting to a path computation element). Conversely, in response to the level not being greater than the threshold, the backup tunnel is allowed to remain activated.