NetGraph

🚧 Work in progress! 🚧

NetGraph is a scenario-based network modeling and analysis framework written in Python. It allows you to design, simulate, and evaluate complex network topologies - ranging from small test cases to massive Data Center fabrics and WAN networks.

Roadmap

• ✅ Fundamental Components: StrictMultiGraph, base pathfinding and flow algorithms
• ✅ Scenario-Based Modeling: YAML-based scenarios with Domain-Specific Language (DSL) describing topology, failures, traffic, and workflow
• ✅ Hierarchical Blueprints: Reusable network templates with nested structures and parameterization
• ✅ JupyterLab Support: Run NetGraph in a containerized environment with JupyterLab for interactive analysis
• ✅ Demand Placement: Place traffic demands on the network with various flow placement strategies (e.g., shortest path only, ECMP/UCMP, etc.)
• ✅ Capacity Calculation: Calculate MaxFlow with different flow placement strategies
• 🚧 Failure Simulation: Model component and risk groups failures for availability analysis with Monte Carlo simulation
• 🚧 Network Analysis: Workflow steps and tools to analyze capacity, failure tolerance, and power/cost efficiency of network designs
• 🚧 Command Line Interface: Execute scenarios from terminal with JSON output for simple automation
• 🚧 Python API: API for programmatic access to scenario components and network analysis tools
• 🚧 Documentation and Examples: Comprehensive guides and use cases
• ❌ Components Library: Hardware/optics modeling with cost, power consumption, and capacity specifications
• ❓ Visualization: Graphical representation of scenarios and results

Status Legend

• ✅ Done: Feature implemented and tested
• 🚧 In Progress: Feature under development
• ❌ Planned: Feature planned but not yet started
• ❓ Future Consideration: Feature may be added later

Quick Start

Docker with JupyterLab (Recommended)

git clone https://github.com/networmix/NetGraph
cd NetGraph
./run.sh build
./run.sh run  # Opens JupyterLab at http://127.0.0.1:8788/

Local Installation

git clone https://github.com/networmix/NetGraph
cd NetGraph
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -e '.[dev]'

Example: Clos Fabric Analysis

from ngraph.scenario import Scenario
from ngraph.lib.flow_policy import FlowPlacement

# Define two 3-tier Clos networks with inter-fabric connectivity
clos_scenario_yaml = """
blueprints:
  brick_2tier:
    groups:
      t1:
        node_count: 8
        name_template: t1-{node_num}
      t2:
        node_count: 8
        name_template: t2-{node_num}
    adjacency:
      - source: /t1
        target: /t2
        pattern: mesh
        link_params:
          capacity: 2
          cost: 1

  3tier_clos:
    groups:
      b1:
        use_blueprint: brick_2tier
      b2:
        use_blueprint: brick_2tier
      spine:
        node_count: 64
        name_template: t3-{node_num}
    adjacency:
      - source: b1/t2
        target: spine
        pattern: one_to_one
        link_params:
          capacity: 2
          cost: 1
      - source: b2/t2
        target: spine
        pattern: one_to_one
        link_params:
          capacity: 2
          cost: 1

network:
  groups:
    my_clos1:
      use_blueprint: 3tier_clos
    my_clos2:
      use_blueprint: 3tier_clos
  adjacency:
    - source: my_clos1/spine
      target: my_clos2/spine
      pattern: one_to_one
      link_count: 4
      link_params:
        capacity: 1
        cost: 1
"""

scenario = Scenario.from_yaml(clos_scenario_yaml)
network = scenario.network

# Calculate maximum flow with ECMP
max_flow = network.max_flow(
    source_path=r"my_clos1.*(b[0-9]*)/t1",
    sink_path=r"my_clos2.*(b[0-9]*)/t1",
    mode="combine",
    flow_placement=FlowPlacement.EQUAL_BALANCED
)
print(f"Maximum flow: {max_flow}")
# Maximum flow: {('b1|b2', 'b1|b2'): 256.0}

Documentation

📚 Full Documentation

• Installation Guide - Docker and pip installation
• Quick Tutorial - Build your first scenario
• Examples - Clos fabric analysis and more
• DSL Reference - Complete YAML syntax
• API Reference - Python API documentation

License

MIT License