Cluster-Based Network Provisioning for Traffic Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network provisioning models, such as the trunk and hose models, face challenges in large-scale networks due to high management complexity and poor bandwidth efficiency, making them unsuitable for large core networks and VPNs.
Innovation Solution
The proposed solution involves partitioning the network into clusters, applying novel node-to-cluster traffic limitations, and using cluster-based trunk or hose models for inter-cluster traffic, allowing for flexible selection of traffic provisioning models within and between clusters, thereby achieving a balance between management complexity and overprovisioning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If the trunk model is used for network provisioning, then bandwidth efficiency is improved, but management complexity increases significantly
Solution Approach 1:
The patent applies segmentation by dividing the network into clusters of nodes. Instead of managing traffic limitations between every pair of nodes (which would require O(N²) parameters), the network is segmented into M clusters, reducing the number of parameters to O(M²). This segmentation maintains bandwidth efficiency by allowing precise control of inter-cluster traffic while simplifying management through hierarchical aggregation.
Solution Approach 2:
The patent introduces a hierarchical dimension to the traditional flat node-to-node provisioning model. By adding the cluster level as an intermediate dimension, the system transforms the parameter complexity from node-pair based (2*(N-1) per node) to cluster-based (fewer parameters per cluster). This dimensional change allows the system to maintain detailed traffic control where needed while aggregating management complexity at higher levels.
2Device complexity
If the hose model is used for network provisioning, then management complexity is reduced, but bandwidth efficiency deteriorates
Solution Approach 1:
The patent applies local quality by allowing different provisioning models to be used in different parts of the network. Intra-cluster traffic can use hose-model-like aggregation for simplicity, while inter-cluster traffic uses more precise trunk-model-like limitations for efficiency. This local differentiation allows the system to optimize for management simplicity in some areas (within clusters) while maintaining bandwidth efficiency in others (between clusters).
3Loss of energy
If node-to-node traffic limitations are enforced in large networks, then bandwidth efficiency is improved, but the number of configuration parameters grows quadratically
Solution Approach 1:
The patent applies merging by combining multiple node-to-node traffic limitations into cluster-to-cluster limitations. Instead of configuring separate parameters for each node pair, the system merges traffic flows at the cluster level, reducing the number of configuration parameters from O(N²) to O(M²) where M << N. This merging maintains bandwidth efficiency by preserving traffic control capabilities while dramatically reducing configuration complexity.
4Loss of energy
If exact traffic matrix information is required for optimal provisioning, then bandwidth efficiency is improved, but measurement and calculation become infeasible in large networks
Solution Approach 1:
The patent applies preliminary action by pre-defining cluster groupings and inter-cluster traffic limitations based on network topology and expected traffic patterns, rather than requiring complete measurement of all node-to-node traffic. This preliminary configuration allows the system to operate effectively with aggregated traffic information at the cluster level, avoiding the infeasible task of measuring and calculating complete traffic matrices in large networks.
Data Source
AI summary
In the area of network provisioning, there is a problem of selecting a suitable traffic-provisioning model for large networks due to the high management complexity of the resource-efficient trunk model and the poor bandwidth efficiency of the easy-to-configure hose model. The invention is based on the idea of partitioning at least part of the network into multi-node clusters, and defining traffic limitations on at least two levels, including the intra-cluster level and the inter-cluster level, where the traffic limitations include one or more node-to-cluster traffic limitations for inter-cluster traffic. Subsequently, cluster-based provisioning of the network is performed based on the traffic limitations. The novel node-to-cluster limitations proposed by the invention are preferably applied in a cluster-based trunk or hose model on the inter-cluster level. In other words, for the description of the inter-cluster traffic (traffic between the clusters) cluster-based trunk or hose models can be used, preferably depending on the available information about the traffic. The cluster-based provisioning makes it possible to find a trade-off between management complexity and overprovisioning.