Overlay Network Dynamic Path Calculation for Data Center Resilience
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cloud computing infrastructure requires over-provisioning and has long setup times due to statically fixed redundancies, making it challenging to ensure high resilience and quality of service, especially in the event of failures within the network.
Innovation Solution
A system comprising a control device and an overlay network that uses segment routing mechanisms, OpenFlow compliance, and connections to multiple domains, allowing dynamic path calculation based on operational needs to reroute data between data centers and adapt to failures, thereby abstracting from Internet service providers' routing policies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If statically fixed redundancies are deployed across multiple data centers, then service resilience is improved, but resource over-provisioning increases and configuration time lengthens
Solution Approach 1:
The patent implements dynamic path selection and resource allocation in the overlay network, allowing routing paths to be adjusted in real-time based on actual traffic conditions and failure states. This replaces static redundancy with dynamic adaptability, maintaining resilience while reducing the need for fixed over-provisioning of resources across all data centers.
Solution Approach 2:
The system changes routing parameters and path selection criteria dynamically based on network conditions, failure detection, and traffic requirements. By modifying routing parameters rather than maintaining fixed redundant paths, the system achieves resilience without permanent over-provisioning of network resources.
2Reliability
If statically fixed redundancies are deployed across multiple data centers, then service resilience is improved, but configuration time lengthens
Solution Approach 1:
The overlay network is pre-configured with multiple possible paths and routing options, but actual path selection is deferred until needed. The control device maintains a pool of pre-established routing rules and path options that can be rapidly activated without requiring lengthy configuration processes when failures occur.
Solution Approach 2:
The system transitions from static pre-configuration to dynamic path selection at runtime. The control device can rapidly reconfigure routing paths in response to failures without requiring lengthy manual configuration processes, achieving both resilience and rapid deployment.
3Adaptability or versatility
If segment routing mechanism with dynamic path calculation is implemented, then adaptability to failures is improved, but control device complexity increases
Solution Approach 1:
The patent introduces a control device as an intermediary that centralizes the complexity of dynamic path calculation and segment routing management. This control device acts as a mediator between the overlay network and underlying infrastructure, handling the computational complexity of failure detection and path recalculation while presenting a simplified interface to network participants.
Solution Approach 2:
The system implements feedback loops where the control device continuously monitors network state, detects failures, and automatically recalculates paths. This feedback mechanism automates the adaptation process, reducing manual intervention complexity while maintaining high adaptability to dynamic failure conditions.
Data Source
Figure 1
AI summary
The invention relates to an overlay network (ON) for a communication network (CN) intended to connect data centres (CSP-A, CSP-B, CSP-C, CSP-D) of a cloud services provider, said overlay network comprising a plurality of particular nodes (NI, N2, N3... N22) configured so as to form paths linking the data centres, and the overlay network being associated with a control device (CC); the particular nodes being further intended to transmit local measurements to the control device intended to calculate a new path within the overlay network and to supply this new path with a set of particular nodes from said plurality, in the event that an anomaly is determined from the local measurements.