Selective Disaggregation Advertisement for Multi-Plane Routing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing routing protocols in data center networks with multi-plane topologies face inefficiencies in disseminating prefix disaggregation notifications, leading to unnecessary propagation to unaffected nodes and slow dissemination, causing traffic disruptions due to link or node failures.
Innovation Solution
An orchestration component selectively identifies affected nodes in a data center fabric and sends targeted disaggregation advertisements using control-plane and data-plane techniques, leveraging Operations, Administration, and Maintenance (OAM) data to notify only affected nodes of failures, allowing them to reroute traffic through alternative communication paths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing routing protocols propagate disaggregation notifications to all leaf nodes, then all nodes receive the notification, but this causes unnecessary traffic disruptions to unaffected nodes and increases network overhead
Solution Approach 1:
The patent segments the set of leaf nodes into affected and unaffected subsets based on the failure location. Instead of notifying all leaf nodes uniformly, the system identifies and notifies only those leaf nodes whose communication paths are actually impacted by the failure, thereby eliminating unnecessary traffic disruptions to unaffected nodes while maintaining routing reliability for affected nodes
Solution Approach 2:
The patent applies local quality by making the notification behavior dependent on the specific failure context. The disaggregation notification is selectively propagated based on the relationship between the failed component and each leaf node's communication paths. This localized approach ensures that notifications are sent only where needed, reducing harmful traffic disruptions while maintaining necessary routing updates
2Reliability
If disaggregation notifications are propagated through the network using existing protocols, then notifications reach all nodes, but the propagation process is slow causing prolonged traffic disruptions
Solution Approach 1:
The patent implements preliminary action by pre-computing and storing the mapping between leaf nodes and their communication paths through spine nodes and intermediate nodes. When a failure occurs, the system can immediately identify affected leaf nodes using this pre-established mapping, eliminating the need for slow network-wide propagation and enabling rapid notification of only the affected nodes
Solution Approach 2:
The patent introduces an orchestration component as an intermediary that maintains the mapping information and performs the selective notification function. This intermediary receives failure notifications, queries the pre-computed mapping to identify affected leaf nodes, and sends targeted disaggregation notifications. This intermediary approach replaces the slow broadcast-based propagation with a faster, query-based selective notification mechanism
3Adaptability or versatility
If the network uses multi-plane topologies with multiple communication paths, then routing flexibility and scalability are improved, but identifying affected nodes becomes more complex
Solution Approach 1:
The patent implements a universal mapping structure that works across multi-plane topologies with multiple communication paths. The mapping between leaf nodes and their communication paths is maintained in a unified format that accommodates the complexity of multi-plane architectures. This universal approach simplifies the identification process despite the underlying routing flexibility and complexity
Solution Approach 2:
The orchestration component serves as an intermediary that manages the complexity of multi-plane topologies. It maintains the mapping information and performs the identification of affected nodes, isolating the complexity from the leaf nodes themselves. This allows the network to benefit from multi-plane routing flexibility while the orchestration component handles the complex identification logic centrally
Data Source
AI summary
Techniques for identifying nodes in a data center fabric that are affected by a failure in the fabric, and selectively sending disaggregation advertisements to the nodes affected by the failure. The techniques include a process where a component monitors the network fabric to identify communication paths between leaf nodes, and determines what leaf nodes would be affected by a failure in those communication paths. The component may detect a failure in the network and determine which communication paths, and thus which leaf nodes, are affected by the failure and send disaggregation advertisements to the affected leaf nodes. In some examples, ingress leaf nodes send data through the fabric that indicate egress nodes for the communication paths. Intermediate nodes along may receive the data from the leaf nodes to identify communication paths, and the notify only affected nodes upon detecting a failure in the network.