Virtualized Network Function Failover via KPI Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wireless network systems face challenges in maintaining high availability and performance due to the complexity of virtualized network functions (VNFs), where failing over one VNF can lead to increased latency and cascading effects, impacting overall end-to-end performance and violating Service Level Agreements (SLAs) if not managed proactively.
Innovation Solution
A Reliability and Availability System (RAS) that uses classification models based on Key Performance Indicators (KPIs) to identify failover conditions and correlate them with remediation models, enabling proactive failover of VNFs to different data centers, ensuring minimal disruption and compliance with SLAs by monitoring KPIs and applying appropriate remediation actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If VNFs are virtualized and co-located on physical machines, then resource utilization and deployment flexibility are improved, but system reliability and availability deteriorate due to cascading failures and latency issues
Solution Approach 1:
The patent segments VNFs into failure domains and implements failure domain isolation, where VNFs are distributed across multiple isolated failure domains (physical machines, data centers, or regions). This segmentation prevents cascading failures by ensuring that a failure in one domain does not propagate to other domains, thereby maintaining system availability while preserving deployment flexibility.
Solution Approach 2:
The patent implements proactive failover mechanisms that anticipate potential failures before they occur. By monitoring KPIs and predicting failures using machine learning models, the system performs failover actions in advance, transitioning VNFs to backup instances in different failure domains before actual failures occur, thus maintaining high availability without compromising deployment flexibility.
2Reliability
If proactive failover is implemented based on KPI monitoring, then system reliability is improved, but system complexity and computational overhead increase
Solution Approach 1:
The patent introduces an intermediary failure domain abstraction layer that simplifies the complexity of proactive failover. Instead of directly managing complex interdependencies between VNFs, the system uses failure domains as intermediaries to organize and isolate VNFs. This abstraction layer provides a standardized interface for failover operations, reducing system complexity while maintaining high reliability through KPI-based monitoring and automated failover execution.
3Reliability
If VNFs are distributed across multiple data centers, then high availability is improved, but latency and inter-site performance issues worsen
Solution Approach 1:
The patent applies local quality by optimizing VNF placement within failure domains to minimize latency. VNFs that require frequent communication are placed in the same or nearby failure domains, while VNFs with independent failure requirements are distributed across different domains. This localized optimization reduces inter-site latency for dependent VNFs while maintaining high availability through geographic distribution, balancing the trade-off between availability and latency.
Data Source
AI summary
A system described herein may provide a technique for the proactive failover of groups of Virtualized Network Functions (“VNFs”) based on monitoring Key Performance Indicators (“KPIs”) associated with the VNFs. KPIs associated with a particular VNF or a set of VNFs may be monitored, and a failover event indicating service degradation or imminent service degradation may be detected based on the KPIs associated with the particular VNF or set of VNFs. Additional VNFs, associated with the particular VNF or set of VNFs, may be further identified for failover from a first data center to a second data center. The additional VNFs may be identified for failover without necessarily monitoring KPIs associated with the additional VNFs, and/or based on factors in addition to KPIs associated with the additional VNFs. For example, the additional VNFs may be failed over based on the KPIs associated with the particular VNF or set of VNFs.