Virtual Switch CPU-Based Traffic Isolation for Multi-Tenant Fairness
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing cloud network forwarding performance isolation methods based on bandwidth and packet rate limitations fail to account for the varying CPU consumption of tenants with different traffic models, leading to unfairness and degradation of network performance for some tenants.
Innovation Solution
A traffic transmission control method that assesses CPU consumption proportions among tenants using historical and current statistical periods to adjust forwarding processing, ensuring fairness and isolation by controlling the virtual switch based on preconfigured CPU consumption proportions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If bandwidth and packet rate limitations are used for forwarding performance isolation, then network resource allocation is simplified, but CPU consumption fairness among tenants with different traffic models deteriorates
Solution Approach 1:
The patent changes the isolation parameter from bandwidth/packet rate to CPU consumption proportion. It introduces a statistical period mechanism that collects CPU consumption data from tenants and calculates their proportion of total CPU usage. The virtual switch adjusts forwarding resource allocation based on this CPU consumption proportion, ensuring that tenants consuming more CPU resources receive proportionally more forwarding capacity, thereby achieving fairness for different traffic models.
2Productivity
If one tenant consumes excessive CPU resources, then virtual switch forwarding performance for other tenants deteriorates, but without CPU-based control mechanisms, resource utilization efficiency decreases
Solution Approach 1:
The patent implements a feedback mechanism where the virtual switch continuously monitors CPU consumption of each tenant during statistical periods. Based on the calculated CPU consumption proportion, the system dynamically adjusts the forwarding resource allocation for the next statistical period. This closed-loop control prevents any single tenant from monopolizing CPU resources while ensuring that high CPU-consuming tenants receive adequate forwarding capacity, thus maintaining both productivity and reliability.
3Speed
If CPU consumption proportion is calculated using only current statistical period data, then response speed is improved, but measurement precision of CPU consumption patterns deteriorates
Solution Approach 1:
The patent applies preliminary action by pre-configuring the statistical period duration and CPU consumption thresholds before operation. During each statistical period, the system accumulates CPU consumption data from all tenants. At the end of each period, it calculates the CPU consumption proportion based on this accumulated data, which provides a more accurate and stable measurement than real-time calculations. This approach balances response speed with measurement precision.
Data Source
AI summary
The present disclosure relates to a traffic transmission control method, an electronic device, and a storage medium. The method includes: obtaining a central processing unit CPU consumption value consumed by each tenant traffic of a cloud network service transmitted via a virtual switch in a current statistical period, to obtain a CPU consumption value corresponding to each tenant; calculating a CPU consumption proportion among tenants based on the CPU consumption value and a CPU consumption value corresponding to each tenant accumulated in a historical statistical period; receiving outbound traffic of each tenant in a next statistical period of the current statistical period; and controlling the virtual switch to perform forwarding processing on the outbound traffic of each tenant in the next statistical period of the current statistical period based on the CPU consumption proportion among the tenants and a virtual switch forwarding performance proportion preconfigured for each tenant.