Roadmap — from shares to shipped MPC

Why this matters

Modules 2–10 stack primitives in a specific order because each later protocol assumes earlier ones as black boxes. Secret sharing (M2) is the representation. Oblivious transfer (M3) is the universal 2PC glue. Garbled circuits (M4) and GMW (M5) are the two major evaluation paradigms — one high-bandwidth constant-round, one low-bandwidth round-heavy. SPDZ (M6) adds malicious security via MACs. Threshold crypto (M7) is MPC applied to key management. PSI (M8) is the most-deployed real-world MPC. Production realities (M9) and hybrid protocols (M10) close the course. The roadmap below names what each module gives you, and what the next module assumes you know.

Demo

Each arrow is 'depends on the ideas from'. Follow it as a reading order if you're skipping around.

M1 (you are here): framework — threat models, real/ideal, additive sharing, OT, Yao-millionaires, when-not-to.
 
   M2 → Shamir (t,n) and Lagrange interpolation
        (generalization of M1's additive sharing)
        |
        ├──→ M5 (GMW arithmetic) — needs Shamir + Beaver triples
        |    |
        |    └──→ M6 (SPDZ) — needs M5 + MACs for malicious security
        |         |
        |         └──→ M7 (threshold crypto) — needs M6's preprocessing + DKG
        |                                       |
        |                                       └──→ M9 (production systems)
        |
        └──→ M8 (PSI) — uses M2 hashes + M3 OT
 
   M3 → OT + OT extension (IKNP)
        (scales M1's single OT to millions)
        |
        └──→ M4 (garbled circuits) — every gate is one OT
             |
             └──→ M10 (hybrid MPC+ZK, MPC+FHE)
 
Every module has the same shape: 8 challenges, 1 project, 1 quiz.
Every course has capstones — MPC has 5, you pick 3 to earn the certificate.
 
Suggested cadence:
  - Days 1-3: M1 (foundations you will reach back to constantly)
  - Week 1:   M2, M3 (cryptographic primitives)
  - Week 2:   M4 or M5 (pick one evaluation paradigm, deeply)
  - Week 3:   M6 (malicious security — this is where real deployments live)
  - Week 4:   M7 or M8 (pick an applied direction)
  - Remainder: M9 + M10, and start a capstone.

Try it yourself

1. Pick which of the 5 capstones sounds most useful for your work. Now trace backwards: which modules does that capstone most depend on? Those are your priority modules.
2. If you can only spend 20 hours on this course, build a personal path through it: list the 3–4 modules you'll do deeply and the 6–7 you'll skim. Defend your pick in two sentences.
3. For each of modules 2 through 10, write one sentence predicting what the central primitive will be. Then check when you get there — it's a good test of your intuition about where crypto lives in each protocol.
4. Find one MPC paper from the last three years that you find interesting (search eprint.iacr.org for 'MPC' or 'multi-party computation'). Identify which module's content it extends. That tells you where the research frontier is.

References

• Evans-Kolesnikov-Rosulek — full TOC
• IACR ePrint — search 'MPC'
• Awesome MPC resource list