SDN Controller Proactive Overload Prevention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing software defined networking (SDN) systems take minutes to tens of minutes to detect and respond to network node failures or congestion, which is insufficient for preventing overload conditions, leading to potential service disruptions and increased maintenance costs due to volatile key performance indicators (KPIs) causing false alarms.
Innovation Solution
A controller node in the communications network monitors key performance indicators of multiple related nodes to compute weighted probabilities of failure, determining when to spawn a new virtual node instance before overload conditions become critical, using a combination of KPIs from the target node and associated nodes to predict and mitigate overload, and utilizing OpenFlow-like protocols to distribute traffic.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional SDN systems monitor and respond to network node failures using conventional methods, then the system can detect failures, but the response time is minutes to tens of minutes which is too slow to prevent overload conditions
Solution Approach 1:
The system performs preliminary actions by continuously monitoring KPIs and computing failure probabilities in advance. When the probability exceeds a threshold, the system proactively spawns a new node instance before the actual failure occurs, enabling prevention rather than just detection of overload conditions
Solution Approach 2:
The system dynamically adjusts the failure detection approach by computing probabilities based on real-time KPI data from multiple nodes. The failure probability threshold and weighting factors are dynamically determined based on the current state of the network, allowing adaptive response to changing conditions
2Measurement precision
If the system monitors KPIs of multiple related nodes and computes weighted failure probabilities to predict overload conditions, then the detection accuracy improves, but the computational complexity and data processing requirements increase
Solution Approach 1:
The system segments the failure prediction process into distinct components: monitoring individual node KPIs, computing individual failure probabilities, weighting based on relationship closeness, and aggregating to determine overall target node failure probability. This modular approach manages complexity while maintaining accuracy
Solution Approach 2:
The system introduces an intermediary computational layer that processes KPI data from multiple nodes and computes weighted failure probabilities. This intermediary layer acts as a mediator between raw KPI data and failure detection decisions, simplifying the overall system architecture while improving prediction accuracy
3Reliability
If a new node instance is spawned immediately when failure probability exceeds threshold, then service continuity is maintained, but the time required to create and advertise the new node (minutes to tens of minutes) may already cause service disruption
Solution Approach 1:
The system performs preliminary failure detection and triggers node provisioning in advance, before the actual failure occurs. By continuously monitoring and computing failure probabilities, the system initiates the node creation process proactively, allowing the new node to be ready before the old node fails, thus avoiding service disruption
Data Source
AI summary
A controller node in a network implementing software defined networking (SDN) proactively detects network overload conditions in a target node and subsequently prevents the conditions from developing by launching a new virtual network node. The detection and prevention are based on the determination of a probability of failure of the target node as a weighted function of information from the target node and probabilities of failure of related nodes.