A proof-of-concept for stateless Bitcoin nodes, based on the accumulator crate. Provided totally as-is and probably will not be maintained, though the authors of this simulation are more than happy to answer your questions.
See the accumulator repo for general setup instructions.
Just cargo run.
The concept for this simulation is adapted from Batching Techniques for Accumulators with Applications to IOPs and Stateless Blockchains (Boneh, Bünz, and Fisch 2018) [Link].
We envision three kinds of actors in a stateless Bitcoin ecosystem (some familiarity with Bitcoin and accumulators is assumed):
- Miners aggregate transactions into blocks, establish consensus on each new block, and store the current chain state in an accumulator (a constant-size aggregation of the UTXO set). Miners publish updates to this accumulator value.
- Users issue transactions from the UTXOs in their possession. When a user presents a UTXO to be spent, they must also present the accumulator witness for that UTXO (proving that the UTXO is unspent with respect to the current accumulator value). In common terminology, users can be understood as light clients.
- Bridges manage witnesses for a number of users, offering these users a liveness guarantee so they don't miss accumulator updates. There are efficient procedures to update a batched set of witnesses, and users can query their individual witnesses on-demand.
In our simulation, each user issues a single transaction per block, and miners cut blocks every t
milliseconds. There are n users assigned to each of m bridge nodes, for a total of n * m users
in the system. Although we include r different miners in the simulation, one of them is always
elected leader to establish consensus.
For more details, please review our code.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent