Module 10: Starlink Network Control, Segment Routing, Automation, and Traffic Engineering

Phase: 3 - Depth Builds on: Modules 03, 08, and 09

Math You’ll Learn

Applied Linear Algebra + Optimization Introduction

Topology work becomes operational when you optimize under capacity, policy, and failure constraints.

• Least squares - fit demand or latency models to noisy telemetry.
  • Starlink application: forecast traffic demand and link utilization.
• Linear programming - maximize/minimize subject to constraints.
  • Starlink application: minimize maximum link utilization across satellite, gateway, POP, and peering links.
• Multi-commodity flow intuition - many source/destination demands share the same network.
• Finite state machines - controller states, rollout states, and failure remediation.
• Control-loop basics - avoid unstable automation that thrashes routes or config.

After this: You can design a topology controller that consumes moving-network state and emits defensible routing or Segment Routing policies.

Resources:

• Strang, least squares chapters
• CLRS graph algorithms and flows
• Service-provider routing references for BGP, IS-IS, MPLS, and Segment Routing

What You’ll Learn

This module replaces generic SDN with the protocols, algorithms, and operational automation that map to Starlink network and topology roles.

Service-Provider Control Plane

• BGP peering/transit, route policy, communities, filtering, max-prefix, and route leaks.
• IS-IS/OSPF as internal routing protocols.
• MPLS, SR-MPLS, SRv6, ECMP, fast reroute, and traffic-engineering policy.
• Control-plane vs data-plane responsibilities in a hybrid space/ground network.
• How gateways, POPs, backbone links, satellites, and laser links become one topology graph.

Topology Service Design

• Inputs: ephemeris, link state, gateway state, POP state, demand matrix, policy, failures, maintenance windows.
• Outputs: next-hop decisions, Segment Routing policies, preferred gateway/POP egress, and capacity reservations.
• Snapshot vs predictive control: current topology, near-future topology, and scheduled changes.
• Route churn control and safe rollout.
• API design with gRPC/Protobuf and REST.

Automation and Operations

• Telemetry: link utilization, latency, loss, route churn, queue depth, and alarms.
• Inventory and desired-state reconciliation.
• Config generation, validation, canary, rollback, and blast-radius control.
• Incident simulation: gateway down, POP isolated, OISL failure, high latency, route leak.
• Linux production operations for network software.

C++ and Python Skills

C++ focus: REST/gRPC service, async orchestration, JSON/Protobuf schemas, Boost.Graph integration, system design.

Python focus: PuLP/SciPy/OR-Tools, NetworkX, telemetry analysis, optimization visualization.

Projects

Project 1: Starlink-Inspired Topology Controller (C++)

Build a simplified network controller.

What you’ll build:

• Ingest topology snapshots from Modules 08 and 09.
• Ingest gateway/POP state and demand matrices.
• Compute paths and generate forwarding decisions or Segment Routing policy objects.
• Expose APIs for route query, link failure, gateway drain, policy update, and recompute.
• Implement failure detection events and automatic rerouting.
• Track route churn and reject unstable updates.

C++ skills used: REST/gRPC, Boost.Graph, async I/O, JSON/Protobuf, system design.

Toolkit: Add TopologyController.

Project 2: Traffic Engineering Optimizer (Python)

Optimize traffic allocation.

What you’ll build:

• Build a network graph with satellite links, gateways, POPs, and peering edges.
• Load a traffic demand matrix between regions/cities.
• Formulate TE as minimizing maximum link utilization subject to capacity and policy constraints.
• Compare shortest path, ECMP, and optimized TE.
• Animate failure and re-optimization.

Python skills used: NetworkX, PuLP/SciPy/OR-Tools, matplotlib.

Technology Reference

Where This Tech Is Used

Books and Resources