SD-WAN Branch Gateway Dynamic Application Migration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In software-defined wide area networks (SD-WAN), migrating applications between uplinks can result in downtime and reduced quality of service due to incomplete context preservation and network health degradation, particularly affecting critical applications like video conferencing and voice over IP.
Innovation Solution
Implementing a dynamic path selection algorithm that classifies applications into criticality classes based on their characteristics, using service level agreements (SLAs) to determine when to migrate applications, and setting migration thresholds to preserve critical applications' quality of service by prioritizing non-critical applications' migration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If applications are migrated between uplinks to maintain quality of service, then network reliability is improved, but downtime and service disruption occur during migration
Solution Approach 1:
The system performs preliminary classification of applications into criticality classes before migration decisions are made. This advance categorization allows the dynamic path selection algorithm to prioritize migrations appropriately, ensuring that non-critical applications are migrated first while critical applications remain on stable uplinks, thereby avoiding downtime for critical services
Solution Approach 2:
The system dynamically changes the migration threshold parameter based on uplink health conditions and application criticality. When uplink health degrades, the algorithm adjusts which applications are subject to migration by referencing their criticality class, ensuring that only non-critical applications are migrated even under degraded conditions, thus preventing service disruption
2Reliability
If dynamic path selection migrates applications to healthier uplinks, then network reliability is improved, but context preservation is incomplete leading to service disruption
Solution Approach 1:
The system applies different migration policies to different applications based on their local characteristics (criticality class). Critical applications are exempted from migration or given higher migration thresholds, while non-critical applications are migrated more freely. This localized quality approach ensures that applications requiring context preservation maintain their connections while still allowing migration for applications where it is safe
3Productivity
If all applications are migrated when uplink health degrades, then load balancing is improved, but critical applications experience quality of service reduction
Solution Approach 1:
The system changes the migration threshold parameter dynamically based on application criticality. Critical applications have higher migration thresholds that prevent their migration until uplink health severely degrades, while non-critical applications have lower thresholds allowing them to be migrated for load balancing. This parameter differentiation enables load balancing without compromising critical services
Data Source
AI summary
An example branch gateway includes processing circuitry and a memory including instructions that cause the branch gateway to perform various functions. The various functions include determining a first uplink health threshold, determining a second uplink health threshold, calculating migration thresholds for a set of non-critical applications, determining that a QoS threshold of a critical application is likely to be imminently breached, selecting a least critical application, based on the migration threshold of the least critical application, and migrating the least critical application from the first uplink to a second uplink.