Network Simulation Segmentation for Data Center Change Testing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional network simulation methods for large-scale data center networks are inefficient due to the complexity of network topology and high computing overhead, and existing systems lack effective means for automated construction of virtual environments, leading to potential network accidents and inability to simulate high-level core devices accurately.
Innovation Solution
A method and apparatus for network simulation that includes running a simulation network comprising full backbone and partial data center devices, using virtual machine images and graph partitioning to reduce cross-physical machine links, and performing network simulation on adjusted networks to test change commands, allowing parallel execution of multiple simulations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If full network simulation is performed for large-scale data center networks, then simulation accuracy is improved, but computing overhead and system complexity increase significantly
Solution Approach 1:
The patent segments the data center network into two parts: core-level network devices (fully simulated) and basic-level network devices (partially simulated). This segmentation allows the system to focus computational resources on critical core devices while reducing simulation complexity for basic devices, thereby resolving the contradiction between simulation accuracy and system complexity.
Solution Approach 2:
The patent extracts and identifies core-level network devices from the overall data center network based on their importance and impact on network performance. By separating core devices from basic devices, the system can apply different simulation strategies to each group, maintaining high accuracy for critical components while reducing overall system complexity.
2Measurement precision
If full network simulation is performed for large-scale data center networks, then simulation accuracy is improved, but computing overhead increases significantly
Solution Approach 1:
The patent applies partial action by simulating only the necessary portions of the network. Specifically, it performs full simulation for core-level devices and partial simulation for basic-level devices, thereby reducing computing overhead while maintaining sufficient simulation accuracy for network change verification.
3Ease of manufacture
If manual construction of simulation environment is performed, then simulation setup is achieved, but time consumption and labor intensity increase
Solution Approach 1:
The patent implements self-service through automated construction of the simulation environment. The system automatically identifies core-level and basic-level network devices, constructs the appropriate simulation topology, and configures simulation parameters without manual intervention, thereby dramatically reducing setup time and labor intensity.
Solution Approach 2:
The patent performs preliminary actions by pre-identifying and categorizing network devices into core-level and basic-level groups before simulation begins. This preliminary classification enables automated environment construction and reduces the time required for simulation setup.
Data Source
AI summary
Embodiments of the disclosure provide a method, an apparatus, a device, and a readable medium for network simulation. The method includes: running a simulation network corresponding to a backbone network and a data center network in a simulation test environment, where the simulation network includes simulation nodes corresponding to a full set of devices in the backbone network and simulation nodes corresponding to a part of devices in the data center network, and the part of devices include a full set of core devices in a core-level network of the data center network and a part of basic devices sampled from a basic-level network; receiving a test request for the data center network, where the test request includes change command for one or more devices; adjusting the simulation network based on the test request; and performing network simulation on the adjusted simulation network in the network simulation environment.