Technical Specification Document

Project Title: Super Alice - Satellite Networking Protocol

Document Version: 1.0
Original Author: Rokas Paulauskas
Co-Authors: Ze Yan Liow, Archit Panigrahi, Sathvika Thorali, Prakash Narasimhan Date Modified: 22/11/2024
Language: C++

---

Execution of Demo

1. To execute demo, download source code extract, proceed to scripts directory and start with: sh run_demo.sh
2. To stop running the demo: sh stop_demo.sh and (command + c) or (ctrl + c) to exit all the running terminals from background.

OS Requirements

1. Unix based OS

Compiler Requirements

1. clang 15 or newer
2. gcc 9.4 or newer

Current use cases

1. Environmental Monitoring and Alert System (Group 14)
2. Soil Health Monitoring in Farms (Group 6)

1. Overview

Super Alice is a peer-to-peer networking protocol designed for communication within Low Earth Orbit (LEO) satellite constellations. Implemented in C++, it leverages a custom UDP protocol optimized for low-latency and lightweight satellite communications. This protocol includes machine learning (ML) for optimized routing and a bootstrap node for peer discovery, enabling efficient data exchange in dynamic and constrained environments.

---

2. Protocol Specifications

2.1 Communication Protocols

• Custom UDP Protocol: Super Alice is built on a custom UDP protocol designed for minimal latency and efficient data transmission in satellite networks. This choice balances speed and simplicity, critical for high-frequency communication in LEO constellations.

2.2 Enhanced Packet Structure

The Super Alice protocol has a modified packet structure to improve flexibility, reliability, and future extensibility. The structure is defined as follows:

#ifndef ALICE_PACKET_HPP
#define ALICE_PACKET_HPP

#include <cstdint>
#include <vector>

namespace alice {

   enum class PacketType : uint8_t {
       HANDSHAKE,
       DATA,
       ACK,
       NACK,
       CONTROL,
       ERROR
   };

   class Packet {
   public:
       uint32_t source_id;
       uint32_t destination_id;
       PacketType type;
       uint8_t priority;
       uint32_t sequence_number;
       uint64_t timestamp;
       uint16_t fragment_id;
       uint16_t total_fragments;
       uint16_t crc;
       uint16_t reserved;
       uint16_t payload_type;
       std::vector<uint8_t> payload;

       Packet(uint32_t source_id, uint32_t destination_id, PacketType type, uint8_t priority, uint32_t sequence_number, const std::vector<uint8_t> &payload);

       [[nodiscard]] std::vector<uint8_t> serialize() const;
       static Packet deserialize(const std::vector<uint8_t> &buffer);
   };
}

#endif

---

3. Field Descriptions

1. Source ID (source_id): 32-bit identifier unique to the originating satellite or ground node.
2. Destination ID (destination_id): 32-bit identifier unique to the target node, ensuring directed communication.
3. Packet Type (type): Enum for packet type:
   • HANDSHAKE: Initializes communication between nodes.
   • DATA: Main data transfer packet.
   • ACK/NACK: Acknowledgments for successful or failed delivery.
   • CONTROL: Command packet for network adjustments.
   • ERROR: Reports errors to aid in diagnostics.
4. Priority (priority): 8-bit field indicating the priority of the packet (0 = low, 255 = high).
5. Sequence Number (sequence_number): 32-bit integer for order in packet transmission.
6. Timestamp (timestamp): 64-bit Unix timestamp for packet timing.
7. Fragment ID (fragment_id): 16-bit identifier for packets that are part of a larger message.
8. Total Fragments (total_fragments): 16-bit field indicating the total number of fragments for reassembly.
9. CRC (crc): 16-bit error correction code for data integrity.
10. Reserved (reserved): 16-bit field for future enhancements.
11. Payload Type (payload_type): 16-bit field to specify the type of payload (e.g., text, binary).
12. Payload (payload): Variable-length data section containing the main content.

---

4. Enhanced Features

4.1 Machine Learning for Routing Optimization

• Purpose: Dynamically adapt routes based on network data, improving latency and throughput.
• Functionality: ML algorithms analyze traffic, node availability, and link quality, adjusting routes as conditions evolve.
• Data: Training data includes positional, latency, and throughput metrics, retrained periodically as new data accumulates.

4.2 Bootstrap Node for Peer Discovery

• Purpose: Streamlines network initialization, connecting new nodes to nearby peers.
• Functionality: The bootstrap node serves as an initial registry for nodes, allowing them to efficiently discover peers and connect.

---

5. Protocol Operation

5.1 Handshake Mechanism

• Uses the HANDSHAKE packet to initiate communication. The bootstrap node aids by introducing nodes to the network.

5.2 Reliable Data Transfer

• DATA packets carry core messages.
• ACK/NACK packets ensure delivery reliability.

5.3 Control Operations

• The CONTROL packet enables real-time adjustments for network configuration and ML-based route updates.

5.4 Error Handling

• ERROR packets report any network issues, allowing rerouting or troubleshooting.

---

6. Routing and Connection Management

• ML-Optimized Routing: ML-driven path predictions enhance data throughput and reduce latency.
• Direct Satellite Routing: Satellite-to-satellite transfers leverage ISLs based on ML-optimized paths.
• Node Failure Resilience: The bootstrap node assists in re-establishing connections when a node fails.
• Load Balancing: Adjusts data flow according to network demand, prioritizing CONTROL and ACK/NACK packets.

---

7. Security and Encryption

Super Alice implements payload encryption, especially for CONTROL and DATA packets, to secure transmissions. Future updates may integrate adaptive ML-driven security mechanisms.

---

8. Future Considerations

1. Enhanced Learning Models: Implement reinforcement learning for more adaptive routing.
2. Advanced Error Correction: Integrate forward error correction for greater robustness.
3. Protocol Switching: Explore the potential for adaptive protocol switching within UDP, leveraging ML to dynamically adjust configurations based on network needs.