Skip to content

Dynamic Simulation

The SymBChainSim is designed to model blockchain networks and consensus protocols under various dynamic conditions. It allows users to configure and observe how different parameters, network behaviors, and workloads impact blockchain performance and resilience.

Core Components

The simulator orchestrates several key components:

  1. Configuration Parameters (Parameters):

    • The entire simulation is driven by a set of parameters loaded primarily from src/Configs/base.yaml. Specific dynamic behaviors and network/workload changes are managed through separate configuration files:
      • src/Configs/behaviour_config.yaml: Defines the number of Byzantine and faulty nodes and their respective behaviors (e.g., sending bad sync data, not responding).
      • src/Configs/dynamic_config.yaml: Dictates if and how network (bandwidth) and workload (transaction rates, sizes) parameters are dynamically updated during the simulation.

  2. The Manager:

    • The Manager (found in src/Simulator/Manager/Manager.py) is the central orchestrator of the simulation. It's responsible for setting up the environment, initializing nodes and the network, and dynamically controlling the simulation's state and parameters during runtime.
    • The Manager uses SystemEvents to introduce changes and trigger actions throughout the simulation.
    • For a detailed explanation of the Manager's functionality, please refer to the Manager documentation.

  3. DynamicSimulation System Events:

    • These are special events, managed by the Manager, that allow the simulation's state and parameters to be updated dynamically during runtime. They are crucial for modeling complex scenarios and behaviors. The Manager dispatches these events to dedicated handlers. Key types of System Events include:
      • Dynamic Simulation: Updates network characteristics (like bandwidth, as seen in src/Simulator/Manager/SystemEvents/DynamicSimulation.py) and workload parameters (like transactions per second and transaction size) dynamically throughout the simulation, sampling from defined distributions in dynamic_config.yaml.
      • Behaviour Events: Manages the introduction of faults (e.g., node crashes) and recoveries, as configured in behaviour_config.yaml.

Simulation Flow (High-Level)

A typical dynamic simulation runs as follows:

  1. Parameter Loading: The Manager loads all initial parameters from base.yaml and other specific configuration files like behaviour_config.yaml and dynamic_config.yaml.
  2. Setup: The Manager initializes the simulation environment, including creating nodes, setting up the network, and scheduling the initial Dynamic Simulation SystemEvents.
  3. Execution Loop: The Manager starts the simulation loop, continuously advancing the Simulation clock and processing events from its queue. During this loop, SystemEvents are triggered to introduce dynamic changes (e.g., updating network conditions, workload generation, simulating node faults etc...).
  4. Termination: The simulation continues until a predefined finish condition is met (e.g., simulation time elapses, a certain number of blocks are produced, or a specific number of transactions are processed).
  5. Results: Upon completion, if enabled, simulation snapshots are saved for post-analysis.

Extending Dynamic Simulation

The above approach allows control over virtually all aspects of the modeled blockchain system. Specifically, anything that is controlled by simulation parameters can be dynamically adapted by introducing new SystemEvent and defining their frequency and effect on such parameters.