Module 04: Variable Links, Handover Resilience, and Transport Behavior

Phase: 2 - Acceleration Builds on: Modules 01, 02, and 03

Math You’ll Learn

Trigonometry Completion + Pre-Calculus Introduction

This module bridges static geometry and time-varying behavior. Starlink links change continuously as satellites move, beams switch, gateways change, and capacity varies.

• Trig review and polar coordinates - antenna patterns and scan angle.
• Angular velocity, arc length, radians - pass duration and handover cadence.
• Functions, composition, and inverses - chain geometry, link quality, and transport behavior.
  • Starlink application: link quality as a function of elevation, weather, beam load, and gateway state.
• Parametric equations - model satellite position and link quality over time.
  • Starlink application: range(t), elevation(t), capacity(t), RTT(t).
• Piecewise functions - model connected, degraded, handover, failed, and recovered states.
• Queueing intuition - backlog accumulation during capacity drops or outages.

After this: You can model variable LEO links and reason about how normal internet transports behave during handover and capacity changes.

Resources:

• Khan Academy - Trigonometry completion and Pre-Calculus
• TCP/IP Illustrated - transport behavior review
• QUIC RFC 9000 and congestion-control references

What You’ll Learn

Deep-space DTN is important background, but it is not the primary Starlink service model. Starlink is an always-on broadband ISP with variable wireless links, rapid handover, gateways, POPs, and normal internet traffic. This module studies how TCP/QUIC behave under those conditions.

Variable LEO Links

• User-terminal satellite handover.
• Gateway handover and POP egress changes.
• Short outages, capacity drops, link degradation, jitter, and packet loss.
• Beam load and congestion as network effects, not only RF effects.
• Difference between disconnected DTN systems and low-latency broadband LEO service.

Transport Behavior

• TCP slow start, congestion avoidance, loss recovery, CUBIC/BBR behavior.
• QUIC path validation, connection IDs, NAT rebinding, and migration concepts.
• Bufferbloat, ACK compression, jitter, and queue management.
• Why high average throughput can still produce bad user experience if latency spikes.
• What to measure: RTT, throughput, packet loss, retransmits, goodput, recovery time.

Test and Emulation

• tc netem for delay, jitter, loss, reorder, and rate limits.
• Synthetic handover events and capacity-change traces.
• Reproducible experiment design: fixed seeds, recorded traces, and clear assumptions.
• DTN/BPv7/LTP as optional contrast for disconnected or store-and-forward systems.

C++ and Python Skills

C++ focus: state machines, POSIX sockets or Boost.Asio basics, serialization, event loops.

Python focus: experiment harnesses, subprocess control, plotting, statistical summaries.

Projects

Project 1: Link-State and Handover State Machine (C++)

Build a reusable state machine for terminal and gateway link state.

What you’ll build:

• Model states: searching, connected, degraded, handover pending, handover executing, failed, recovered.
• Define events: elevation below threshold, better satellite available, packet loss spike, gateway unavailable, timeout, recovery.
• Record transition history with timestamps and reason codes.
• Feed synthetic link-quality traces into the state machine.
• Export event logs as JSON for analysis.

C++ skills used: enums, state machine design, classes, callbacks, unit tests.

Toolkit: Add HandoverStateMachine.

Project 2: TCP/QUIC Handover Experiment (Python)

Compare transport behavior across Starlink-like link events.

What you’ll build:

• Use tc netem or a pure simulator to model delay, loss, jitter, variable capacity, and short outages.
• Transfer data over TCP and, if practical, QUIC tooling.
• Inject handover-like events and capacity drops.
• Plot throughput, RTT, packet loss, retransmits, and recovery time.
• Write a results note explaining when the transport recovered well and when it did not.

Python skills used: socket tooling, subprocess, timing, pandas optional, matplotlib.

Technology Reference

Where This Tech Is Used

Books and Resources