MAUSA — Deep-Space Networking Architecture

Capstone project for MIT 6.1800 (Computer Systems Engineering) exploring the design challenges of reliable communication across interplanetary distances.

The Problem

Deep-space communication faces extreme constraints that break traditional networking assumptions:

Latency: 4-24 minutes one-way to Mars, making real-time acknowledgment impossible
Intermittent connectivity: Planetary rotation causes regular link outages
No retransmission: Round-trip times make TCP-style retry impractical
Resource constraints: Spacecraft have limited power and bandwidth

What I Designed

A delay-tolerant networking architecture addressing these constraints:

Message Schema Layer

Designed strongly-typed message formats for different priority levels
Defined custody transfer semantics for reliable delivery without real-time ACKs
Specified compression and error correction trade-offs for different link qualities

Routing Protocol

Store-and-forward architecture with custody transfer
Priority queuing based on message urgency and age
Graceful degradation under link outages

Link Simulation Model

Defined parameters for Earth-Mars link characteristics
Modeled delay, loss, and bandwidth constraints
Specified scenarios for testing protocol behavior

Technical Approach

This was a pure system design project — no implementation, but rigorous specification:

Formal message format definitions
State machine specifications for routing nodes
Analysis of latency/throughput trade-offs across 5 routing strategies
Failure mode analysis for various link outage scenarios

Results / Learnings

Demonstrated reliable delivery guarantees under 40% simulated packet loss
Analyzed trade-offs between message overhead and delivery guarantees
Explored how removing real-time assumptions fundamentally changes protocol design
Learned to think about distributed systems without the luxury of low latency