Multi-layer Network Resiliency via Server-Client Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multi-layer network designs face challenges in optimizing resiliency between the server and client layers, leading to non-optimal restoration path selection and potential partitioning of the client layer due to single failures in the server layer, which complicates network economics and reduces the originally designed resiliency.
Innovation Solution
The method involves determining a minimal spanning tree in the client layer and a Steiner tree in the server layer, adding resiliency paths to the client layer based on potential failures in the Steiner tree to avoid partitioning, and utilizing exclusion list criteria for routing in the server layer to maintain maximal diversity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual network design processes are used to provision protection bandwidth in multi-layer architecture, then network resiliency is achieved, but the design becomes complicated and non-optimal as network complexity increases
Solution Approach 1:
The system performs preliminary analysis of the server layer topology and potential failure points before client layer design is finalized. By pre-determining the minimal spanning tree and identifying critical server layer links, the system enables optimal client layer resiliency design without complex manual intervention during deployment.
Solution Approach 2:
The system establishes a feedback loop where server layer topology changes automatically trigger client layer resiliency re-evaluation. The controller continuously monitors server layer state and adjusts client layer protection paths accordingly, eliminating the need for complex manual redesign when network topology evolves.
2Ease of operation
If server layer control plane functions operate independently without considering the overlay network, then server layer operations are simplified, but client layer resiliency erodes due to non-optimal restoration path selection
Solution Approach 1:
The controller acts as an intermediary between the server layer control plane and client layer operations. It receives server layer topology information, analyzes potential failures, and provides guidance for optimal client layer protection path selection, enabling coordinated multi-layer resiliency without compromising server layer operational simplicity.
Solution Approach 2:
The system adds a new dimension of control by introducing a dedicated controller that operates above both server and client layers. This controller encompasses knowledge of both layers and makes resiliency decisions that consider the entire multi-layer architecture, rather than optimizing each layer independently.
3Quantity of substance
If multiple client layer links are physically co-routed in the server layer, then network resource utilization is improved, but a single server layer failure causes multiple client layer failures
Solution Approach 1:
The system segments client layer links into different protection groups based on their server layer physical routes. By identifying links that share common server layer infrastructure, the system can assign different protection strategies to different segments, ensuring that a single server failure does not cascade to multiple client layer failures while still allowing resource sharing where safe.
Solution Approach 2:
The system dynamically changes the protection parameter (protection path selection) for client layer links based on server layer topology analysis. Links that are physically co-routed in the server layer are assigned different protection paths in the client layer, transforming the risk profile from correlated failures to independent failure modes while maintaining resource efficiency.
Data Source
AI summary
Systems and methods for path computation of a service in a multi-layer network including a client layer and a server layer include determining correlations between the client layer and the server layer; assigning data to one of client layer links and server layer links based on diversity between one another responsive to the determined correlations; and determining a resiliency path for a service from a current path using the data to determine diversity between the resiliency path and the current path in the client layer and the server layer.