Controller Placement in Split Architecture Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In split architecture networks, the placement of controllers is arbitrary and lacks a systematic approach, leading to increased deployment costs, wiring complexity, and reduced network resilience due to the separation of control and data planes, which existing solutions fail to address effectively.
Innovation Solution
A method and system for determining optimal controller placement by graphing the network topology, applying agglomerative or partitive clustering processes to minimize inter-cluster link lengths, and assigning controllers to central nodes within clusters, thereby reducing disruption from link, switch, or connectivity failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If controllers are placed arbitrarily in split architecture networks, then deployment is simplified, but network resilience and performance deteriorate due to increased disruption from link or switch failures
Solution Approach 1:
The patent applies preliminary action by pre-calculating optimal controller placements using clustering algorithms (K-means, hierarchical clustering) before network deployment. The system computes cost functions based on topological metrics (shortest path distances, betweenness centrality) and determines optimal locations in advance, eliminating the need for arbitrary placement while ensuring network resilience from the start.
Solution Approach 2:
The patent employs parameter changes by optimizing controller placement based on multiple topological parameters including shortest path distances, betweenness centrality, and clustering coefficients. The system adjusts placement decisions by evaluating different parameter combinations and selecting configurations that minimize disruption cost functions, thereby improving network resilience through data-driven parameter optimization.
2Reliability
If controllers are placed to maximize network resilience, then disruption from failures is reduced, but deployment complexity and wiring complexity increase
Solution Approach 1:
The patent applies self-service by implementing automated algorithms that independently calculate optimal controller placements without requiring manual network analysis or complex configuration procedures. The system uses self-contained optimization routines that take network topology as input and automatically generate placement recommendations, reducing deployment complexity while maintaining high network resilience.
Solution Approach 2:
The patent replaces manual, mechanical deployment processes with computational algorithms. Instead of relying on network administrators to manually analyze topology and determine optimal placements, the system uses automated clustering algorithms and cost function evaluations to substitute human analysis with computational optimization, thereby reducing deployment complexity.
3Productivity
If controllers are placed centrally to minimize inter-cluster link lengths, then network performance improves, but vulnerability to single points of failure increases
Solution Approach 1:
The patent applies segmentation by dividing the network into multiple clusters, each with its own centrally-optimized controller placement. The clustering algorithms (K-means, hierarchical clustering) partition the network topology into distinct groups, allowing each cluster to have localized optimal placements that improve performance while distributing failure risks across multiple segments rather than concentrating control in a single point.
Solution Approach 2:
The patent employs local quality by optimizing controller placements independently for each network cluster rather than using a uniform global placement strategy. The cost function evaluation and placement optimization are performed locally within each cluster based on its specific topological characteristics, enabling tailored placements that improve local performance while maintaining overall network resilience through distributed architecture.
Data Source
AI summary
A network topology design system to determine placement of a set of controllers within a network with a split architecture, the placement of the set of controllers selected to minimize disruption of the split architecture network caused by a link failure, a switch failure or a connectivity loss between the set of controllers and the data plane components. The system performs a method including graphing a topology of the split architecture network, determining a set of clusters of nodes within the graph by applying an agglomerative clustering process or a partitive clustering process, determining, a centroid for each cluster in the set of clusters, assigning one of the set of controllers to each network element corresponding to a determined centroid in the graph, and assigning each controller to control a set of network elements corresponding to a cluster in the graph.