Research Python Simulation Libraries

[Teagueanator]

Figure out and write an explanation of what simulation tools we should use for representing:

1. Nodes
2. Messages
3. Distance
4. What Data structures

[victoralien]

I tried to keep it simple, but for our discrete sim we also might be able to use SimPy

1. Nodes

What to represent:

• IoT sensors (>=70% stationary)
• Mobile clients (0.5-2 m/s movement)
• MQTT brokers
• Gateways

Data Structure:

class Node:
    id: str
    type: NodeType  # SENSOR, MOBILE_CLIENT, BROKER, GATEWAY
    position: (x, y)  # coordinates on map
    state: NodeState  # SLEEP, AWAKE, CONNECTED, DISCONNECTED
    energy_level: float
    wireless_tech: WirelessType  # BLE, WIFI, ZIGBEE
    
    # For mobile nodes
    velocity: (vx, vy)  # movement vector
    mobility_model: MobilityModel  # GRID, RANDOM_WAYPOINT

Implementation Tool: Use object-oriented classes with inheritance like SensorNode, MobileNode, BrokerNode, ect

2. Messages

What to represent:

• MQTT publish/subscribe messages
• QoS levels (0, 1)
• Message duplicates and retransmissions
• Keep-alive messages
• Last Will Testament (optional)

Data Structure:

class MQTTMessage:
    message_id: int
    topic: str
    payload: bytes
    qos: int  # 0 or 1
    retain: bool
    duplicate: bool
    timestamp: float
    sender_id: str
    receiver_id: str
    
    # Tracking
    delivery_attempts: int
    acknowledged: bool
    end_to_end_latency: float

Implementation Tool: Use queues queue.Queue or collections.deque for message buffers and broker queues

3. Distance

What to represent:

• Physical distance between nodes
• Signal propagation and range limits
• Coverage areas for different wireless technologies

Data Structure:

class DistanceModel:
    # Simple Euclidean distance
    def calculate_distance(node1_pos, node2_pos):
        return sqrt((x2-x1)^2 + (y2-y1)^2)
    
    # Range limits by technology
    BLE_RANGE = 100  # meters
    WIFI_RANGE = 250  # meters  
    ZIGBEE_RANGE = 75  # meters
    
    def is_in_range(node1, node2):
        distance = calculate_distance(node1.position, node2.position)
        return distance <= get_range_limit(node1.wireless_tech)

Implementation Tool: Use 2D coordinate system with simple geometric calculations

4. Data Structures Summary

Core Simulation Framework

• Event Queue: heapq or PriorityQueue for discrete event simulation
• Time Management: Single global clock with scheduled events
• Node Registry: dict mapping node IDs to node objects

Network Topology

• Adjacency Lists: Track which nodes can communicate with each other
• Coverage Maps: 2D grids showing signal coverage areas
• Broker Connections: Track client-broker relationships

Message Handling

• Per-node Queues: Separate transmit/receive queues for each node
• Broker Topics: dict mapping topics to subscriber lists
• Session State: For persistent sessions when clean_session=false

Mobility & Events

• Movement Scheduler: Queue of position update events for mobile nodes
• Topology Events: Scheduled broker failures, gateway relocations
• Statistics Collectors: Running totals for delivery ratio, latency, energy consumption

[Canyon-C]

I looked into both SimPy and ns-3 (a more feature-rich C++ package, usable with Python bindings), and it seems as though ns-3 provides more of a foundational structure for our needs. As SimPy isn't inherently designed for network simulations, it lacks modules and abstractions for many common network-related protocols (MQTT). ns-3 has immense support for a diverse set of models and is something we should consider. Additionally, with SimPy, we would be required to design and develop lots of our own classes and data structures (which isn't inherently bad), just another thing we have to worry about.