VLAN Routing via Packet Encapsulation and Hash Classes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Data center networks face challenges in managing millions of communication flows with complex fabric topologies while maintaining a limited state, where conventional ensemble routing techniques sacrifice fine-grained control for individual flows and struggle with handling large numbers of device addresses, leading to non-optimal paths and traffic congestion.
Innovation Solution
The implementation of packet encapsulation with additional header fields that identify edge devices for handling packets, using a hash class and traffic class to optimize routing on virtual local area networks (VLANs), allowing for improved traffic management and load balancing by encapsulating packets at the edge of the data center network and selecting routing VLANs based on traffic measurements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional ensemble routing techniques are used to manage millions of communication flows, then the network can handle large numbers of flows, but fine-grained control for individual flows is sacrificed and non-optimal paths are selected
Solution Approach 1:
The patent segments the routing control into two levels: ensemble-level routing for macro traffic management and packet-level encapsulation with hash class/traffic class for fine-grained control. This segmentation allows the system to maintain both high throughput for millions of flows and precise control for individual flow optimization
Solution Approach 2:
The patent introduces encapsulation headers with hash class and traffic class fields as intermediaries between the packet and the routing decision mechanism. These intermediary fields enable detailed flow classification and routing optimization without requiring complex per-flow state in core switches
2Device complexity
If conventional ensemble routing is used, then routing state information is limited, but traffic congestion occurs and load balancing is insufficient
Solution Approach 1:
The patent performs preliminary classification of packets into hash classes and traffic classes at the ingress point before routing decisions are made. This preliminary action enables efficient load balancing and congestion avoidance downstream without requiring complex real-time routing state management in core switches
Solution Approach 2:
The patent changes the routing parameters by introducing hash class and traffic class fields in the encapsulation header. These parameter changes enable fine-grained traffic classification and routing optimization while maintaining limited state in core switches through hash-based routing
3Productivity
If packet encapsulation with hash class and traffic class is implemented, then routing optimization and traffic de-congestion are improved, but processing complexity at edge devices increases
Solution Approach 1:
The patent implements self-service by having edge devices automatically perform hash class and traffic class classification based on packet characteristics. This self-service approach reduces the need for complex external control mechanisms while improving routing optimization and traffic de-congestion
Data Source
AI summary
A system and method are provided to route packets in a data center network. Individual packets are encapsulated at an edge of the data center network, so that each encapsulated packet includes a set of header fields, such as a tenant identifier. For each encapsulated packet, a hash class is determined from the set of header fields. A routing virtual local area network (VLAN) is selected for the packet based on the tenant identifier and the hash class.