SD-WAN Cluster Anchoring via Dynamic Bucketmaps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In software-defined wide area networks (SD-WAN), single network infrastructure devices often fail to provide the required availability and capacity to handle the load, leading to potential failures and service disruptions, necessitating a solution to enhance high availability and capacity.
Innovation Solution
Clustering multiple network infrastructure devices together, with a leader distributing load and using bucketmaps to dynamically assign new devices to anchor controllers, and a network orchestrator monitoring health parameters to proactively adjust tunnel configurations and failover strategies to ensure high availability and capacity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple network infrastructure devices are clustered together to handle load, then availability and capacity are improved, but device complexity increases
Solution Approach 1:
The system segments the cluster management functionality by introducing a leader device that coordinates anchoring decisions. New anchored devices are assigned to specific branch gateways based on a distributed hash table (DHT) structure, where the leader device partitions the namespace and assigns ranges to follower devices. This segmentation allows the cluster to scale while maintaining manageable complexity at each node.
Solution Approach 2:
The anchoring system implements dynamic load distribution where the leader device can reassign anchoring ranges to follower devices based on current cluster conditions and device capabilities. This dynamic reassignment allows the system to adapt to changing loads and maintain optimal performance without requiring complete reconfiguration of the cluster architecture.
2Productivity
If load is distributed across multiple branch gateways, then capacity is improved, but coordination complexity increases
Solution Approach 1:
The leader device serves as an intermediary that mediates all anchoring decisions in the cluster. When a new anchored device joins, it contacts the leader device which then determines the appropriate follower branch gateway based on the DHT namespace partitioning. This intermediary approach centralizes coordination logic while maintaining distributed load handling, reducing the coordination burden on individual follower devices.
Solution Approach 2:
The system performs preliminary actions by pre-establishing the DHT namespace partitioning and assigning anchoring ranges to follower devices in advance. This preliminary configuration allows follower devices to immediately handle anchoring requests for their assigned namespaces without requiring real-time coordination with the leader, reducing operational complexity during normal load distribution.
3Reliability
If health parameters are monitored to enable proactive failover, then reliability is improved, but measurement and detection difficulty increases
Solution Approach 1:
The leader device implements continuous feedback monitoring of follower device health parameters including CPU utilization, memory usage, and operational status. Based on this feedback, the leader device can proactively reassign anchoring ranges from unhealthy followers to healthy ones, or trigger failover procedures when critical thresholds are exceeded. This feedback mechanism enables reliable automatic adaptation to device failures without complex manual intervention.
Data Source
AI summary
An example network orchestrator of a SD-WAN is configured to classify, based on operating health information, each branch gateway of a cluster in an operating health class of a set of operating health classes. A bucketmap is generated based on the classifications of each branch gateway of the cluster. The bucket map may be transmitted to a leader branch gateway of the cluster. Each branch gateway may be reclassified in an operating health class. An updated bucketmap may be transmitted to a leader branch gateway.