Flyclient: Super-Light Clients for Cryptocurrencies
summary
Proposal: a super-light client called Flyclient that replaces linear header storage with a Merkle Mountain Range committing cumulative block weights, uses probabilistic sampling plus proofs-of-proof-of-work to decide between competing chains, and transforms interactive checks into non‑interactive proofs via Fiat‑Shamir. Claims: proof sizes drop from ~2.2 GB of headers to about 1–3 KB, soundness holds under up to a specified adversarial mining fraction, and the design handles variable difficulty by storing leaf weights. Deployment options: hard fork, soft fork, or velvet fork where occasional MMR commits suffice.
AI-written summary · about
chapters (13)
- 0:00Stanford University
- 1:08Blockchain size: A growing problem
- 1:36Simple Payment Verifying Cient (Satoshi 2008)
- 2:27Use the longest chain rule
- 2:58Can verify specific transactions with help
- 3:37SPV Properties and Problems
- 7:59Merkle Mountain Ranges Todd 16
- 9:54Flyclient: A different approach to super light clients
- 13:37Flyclient Strawman 1 problem: Forking
- 13:59Flyclient idea: Find Fork Point
- 16:15Sampling distribution of randomized probacol
- 18:25Optimizing Distribution
- 21:03Evaluation (Ethereum sorry)
related talks
54mSharp Forks & Articmine on Monero’s Adaptive Blockweight
31mTowards making private telemetry as ubiquitous as TLS
01h 01mKey transparency: Blockchain meets NoiseSocket
25mMobileCoin Fog: a cloud you can't see through
01h 02mCrosslink and Possible Futures of Proof-of-Stake in Zcash
17mThe Pod: Post-Quantum Readiness
25mEgalitarian Computing
37mCryptocurrency - building a wallet on your phone