Dual-WAN Link Selection Using Application-Server Latency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network management devices struggle to accurately manage latency when routing communications to application servers over wide area networks due to high variability in latency depending on the location of application servers and the route through the WAN, leading to suboptimal performance for specific applications.
Innovation Solution
A network management device dynamically evaluates latency over multiple WAN links for specific application servers, selecting the link with the lowest latency for routing traffic, considering factors like bandwidth capacity and application characteristics to ensure efficient and low-latency communication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If static host monitoring is used to test WAN links, then latency reduction is achieved across a plurality of applications on average, but poor latency management is provided for some specific applications
Solution Approach 1:
The patent segments the latency measurement process by creating separate latency profiles for different application servers rather than using a single static host monitoring approach. This allows the system to measure and manage latency specifically for each application server that client devices need to access, resolving the contradiction between average performance and specific application management.
Solution Approach 2:
The network management device performs preliminary latency measurements to multiple application servers before client devices need to access them. By pre-determining latency characteristics for specific application servers and storing this information, the system can make informed routing decisions without real-time measurement delays, achieving both average latency reduction and specific application optimization.
2Adaptability or versatility
If WAN link latency is determined for specific application servers, then sophisticated latency management choices are enabled, but computational and bandwidth requirements in the network management device increase
Solution Approach 1:
The system performs latency measurements to application servers in advance and stores the results as latency profiles. This preliminary action allows the network management device to make sophisticated routing decisions based on pre-collected data, avoiding the need for continuous real-time measurements and reducing ongoing computational and bandwidth requirements.
Solution Approach 2:
The patent creates latency profile copies for different application servers that can be referenced during routing decisions. Instead of performing complex real-time calculations for each routing decision, the system uses pre-generated latency profiles as references, significantly reducing the computational burden while maintaining sophisticated latency management capabilities.
3Reliability
If the network management device selects WAN links based on latency for critical applications, then user experience is enhanced, but network complexity increases
Solution Approach 1:
The patent segments applications into different categories (e.g., real-time critical applications versus non-critical applications) and applies different latency management strategies to each segment. This allows the system to enhance user experience for real-time applications through sophisticated latency-based routing while using simpler strategies for non-critical applications, thereby managing overall network complexity.
Solution Approach 2:
The system applies different quality levels of latency management to different applications based on their requirements. Critical real-time applications receive sophisticated latency-based routing with detailed latency profiles, while non-critical applications use simpler routing strategies. This local quality approach enhances user experience where needed without unnecessarily increasing complexity across the entire network.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A network management device for enabling client devices to access services provided over a wide area network, WAN, and a method for network management device are provided. The network management device comprises first and second WAN links and is configured to determine IP address data including an IP address associated with an application server. Latency indications associated with the application server are determined using both the first and second WAN links, and traffic to the application server is routed over one of the first or second WAN links based on a comparison of their respective latency indications.