Relaxed Constrained Shortest Path First Algorithm
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Constrained Shortest Path First (CSPF) algorithms in computer networks use hard constraints, which do not allow for tradeoffs between attributes, leading to high path costs and inefficiencies, and can result in no available paths if constraints are not met, requiring cumbersome manual relaxation of constraints.
Innovation Solution
The Relaxed Constrained Shortest Path First (R-CSPF) algorithm configures a target bandwidth and boundary constraints on a path computation device to compute paths within acceptable bandwidth and cost limits, selecting paths that are closest to these constraints, and allows for weighting factors to prioritize cost or bandwidth violations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If hard constraints are used in CSPF path computation, then constraint satisfaction is improved, but path cost increases and path availability decreases
Solution Approach 1:
The patent transforms static hard constraints into dynamic soft constraints that can be adjusted during path computation. The system dynamically determines constraint violations and applies penalty costs rather than completely excluding links, allowing the path selection to adaptively balance constraint satisfaction with overall path optimality.
Solution Approach 2:
The patent changes the parameter representation of constraints from binary (satisfied/not satisfied) to continuous (violation degree). By measuring constraint violations as quantitative values and incorporating them into the path cost function, the system enables gradual optimization rather than abrupt constraint enforcement.
2Reliability
If hard constraints are used in CSPF path computation, then constraint satisfaction is improved, but path availability decreases
Solution Approach 1:
The system dynamically adjusts constraint enforcement based on network conditions and computation results. Rather than rigidly excluding all links that don't meet constraints, the system dynamically calculates violation penalties that allow some flexibility while maintaining overall constraint satisfaction goals.
Solution Approach 2:
The patent accepts temporary or partial constraint violations as acceptable trade-offs when they lead to significantly better overall path solutions. The system is willing to tolerate minor, short-lived constraint breaches if they result in substantially improved path metrics.
3Productivity
If manual constraint relaxation is performed, then path availability is improved, but configuration complexity increases
Solution Approach 1:
The system performs automatic constraint relaxation and optimization without requiring manual administrator intervention. The path computation device autonomously calculates violation penalties, evaluates trade-offs, and selects optimal paths based on configured objectives, eliminating the need for manual constraint adjustment.
Solution Approach 2:
The system implements feedback loops where path computation results inform subsequent optimization decisions. The system continuously evaluates constraint violations and adjusts path selection based on measured performance, automatically refining solutions without manual reconfiguration.
4Device complexity
If tradeoffs between constraints are not considered, then constraint enforcement is simplified, but solution optimality decreases
Solution Approach 1:
The patent merges multiple constraint evaluations into a unified path cost function. Instead of treating each constraint separately, the system combines violation penalties for bandwidth, latency, and other constraints into a single composite metric that guides path selection, simplifying the overall computation while capturing trade-offs.
Solution Approach 2:
The system transforms multiple constraint parameters into a unified optimization framework where violation degrees are quantified and weighted. By changing from discrete constraint checks to continuous parameter optimization, the system achieves more precise solution evaluation.
Data Source
AI summary
In one embodiment, a target bandwidth, a lower boundary constraint on bandwidth, and an upper boundary constraint on cost are configured on a path computation device. The device may then compute a set of paths that is within the lower bandwidth boundary and upper cost boundary constraints. If the set of paths contains one or more paths, the device may select a path having a bandwidth value that is closest to the target bandwidth. However if the set of paths contains no paths, the device may search amongst paths that violate either the maximum cost or minimum bandwidth limits, selecting the path closest to the violated constraint (e.g., whose combined violation is minimized). Also, a weighting factor may be configured to establish the relative importance for a cost violation in comparison to a bandwidth violation.