Link Aggregation Group Synchronization Ports for Virtual Machine Traffic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In virtualized datacenter environments, load balancing across virtual machines within a single link aggregation group is inefficient due to the reliance on a virtual load balancer, which acts as a choke point for network throughput, limiting the distribution of workload across multiple physical or virtual machines.
Innovation Solution
Configuring a link aggregation group (LAG) across multiple host devices, where synchronization ports are used to transmit traffic between virtual machines on different host devices, excluding them from shared LAGs with other host devices, and employing a switch to perform load balancing between ports within the LAG, thereby distributing the load balancing function across multiple host devices and reducing the dependency on a single virtual load balancer.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a virtual load balancer is used to distribute workload across multiple physical or virtual machines, then load balancing functionality is provided, but network throughput is limited due to the choke point effect
Solution Approach 1:
The patent segments the load balancing function by introducing dedicated synchronization ports that are separate from the LAG ports. These synchronization ports handle control traffic between host devices independently, allowing the LAG ports to focus on data traffic without the overhead of virtual load balancer processing. This segmentation resolves the contradiction by eliminating the choke point while maintaining load balancing capabilities.
Solution Approach 2:
The patent introduces synchronization ports as intermediary elements that facilitate communication between host devices sharing a LAG. These ports act as mediators for control traffic, enabling the switch to perform load balancing decisions without requiring all traffic to pass through a virtual load balancer. This intermediary mechanism resolves the throughput limitation while preserving the load balancing function.
2Productivity
If all traffic passes through a virtual load balancer, then centralized load balancing control is achieved, but workload distribution efficiency is reduced
Solution Approach 1:
The patent segments traffic processing into two paths: control traffic through synchronization ports and data traffic through LAG ports. This segmentation allows workload distribution decisions to be made at the switch level without requiring all traffic to traverse the virtual load balancer, thereby improving workload distribution efficiency and reducing processing delay.
Solution Approach 2:
Instead of having all traffic pass through the virtual load balancer for centralized control, the patent inverts the approach by enabling direct traffic flow between host devices through the switch, with load balancing decisions made at the network infrastructure level. This inversion eliminates the choke point and improves both efficiency and speed.
3Productivity
If synchronization ports are excluded from shared LAGs, then traffic between virtual machines on different host devices can be transmitted directly, but LAG port utilization is reduced
Solution Approach 1:
The patent segments ports into two distinct categories: synchronization ports for control traffic and LAG ports for data traffic. This segmentation allows direct traffic transmission between virtual machines on different host devices through synchronization ports while maintaining LAG port utilization for high-speed data transfer. The separation resolves the contradiction by enabling both direct transmission and efficient port utilization without configuration complexity.
Data Source
AI summary
Exemplary methods, apparatuses, and systems include a first host device determining that a first packet from a first virtual machine (VM) within the first host device is to be transmitted to a second VM on a second host device and that the first host device and the second host device each transmit or receive packets via ports within a first link aggregation group (LAG). In response to determining that the first host device and the second host device each transmit or receive packets via ports within the first LAG, the first host device transmits the first packet from a first synchronization port of the first host device to a second synchronization port of the second host device. The first and second synchronization ports are excluded from sharing a common LAG with any ports of another host device.