Distributed Load Balancing in SDN Network Virtualization Edge
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional load balancers in software-defined networks (SDNs) introduce performance restrictions and prolonged transmission delays due to their architecture, which is sub-optimal for large-scale distributed applications, as they require multiple traversals through the underlay network, leading to inefficiencies and increased end-to-end delays.
Innovation Solution
Implementing a distributed load-balancing method within the Network Virtualization Edge (NVE) of an SDN, where a load-balancer instance is instantiated and configured to select a real server from a pool based on packet hashing, reducing the need for external load-balancers and integrating load-balancing with NVE forwarding operations, thereby minimizing additional computing resources and transmission delays.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional load balancers are used in SDN, then load-balancing function is provided, but performance restrictions and prolonged transmission delays occur
Solution Approach 1:
The patent merges the load-balancing function with the NVE forwarding operations by integrating the load-balancer instance directly into the NVE. This consolidation eliminates the need for separate load-balancer devices and reduces the number of network traversals required, thereby maintaining load-balancing functionality while reducing transmission delays.
Solution Approach 2:
The patent segments the load-balancing function from the traditional external load-balancer architecture and distributes it across multiple NVE instances. Each NVE maintains a local load-balancer instance that can independently make forwarding decisions, eliminating the bottleneck of centralized load-balancing and reducing end-to-end delays.
2Productivity
If external load-balancers are used, then load distribution is achieved, but additional computing resources and device complexity increase
Solution Approach 1:
The NVE is designed to perform multiple functions simultaneously, including both forwarding operations and load-balancing. By making the NVE universal and capable of handling both tasks, the patent eliminates the need for dedicated external load-balancer devices, thereby reducing device complexity and computing resource requirements while maintaining effective load distribution.
Solution Approach 2:
Each NVE instance autonomously performs load-balancing decisions for its own traffic using its integrated load-balancer instance. This self-service approach eliminates the need for external load-balancer devices to process traffic, reducing overall system complexity and computing resource requirements while maintaining effective load distribution across the network.
Data Source
AI summary
A load-balancer instance is instantiated in a network virtualization edge (NVE) in a software defined network (SDN). A forwarding table in the NVE is modified to indicate that a next hop for a packet having a destination address of the load-balancer instance is to be resolved by the load-balancer instance. From a portion of the packet, and using the load-balancer instance, a value usable to select a singular next hop to a first real server in a pool of real servers managed by the load-balancer instance is determined. The packet is forwarded, using the modified forwarding table, the packet through an underlay of the SDN such that the packet tunnels from the NVE to a first NVE, the second NVE hosting the first real server.