CompactChain

An efficient UTXO-based blockchain

[Paper]

Annual power consumption for maintaing bitcoin network is more than electricity consumption of country Sweden [1].

UTXO-based blockchains like bitcoin search in large memory pool for verifying “unspent transactions”, due to evergrowing UTXO size demands more memory size which will be infeasible in upcoming future.

We proposed a new architecture which eliminates the storage of UTXO set rather we use RSA accumulators. As Boneh proposed in “Batching Techniques for Accumulators with Applications to IOPs and Stateless Blockchains”, we can dynamically add and delete elements in RSA accumulator with computational complexities of \(\mathcal{O}(n)\) and \(\mathcal{O}(n^2)\).

Delete operation with \(\mathcal{O}(n^2)\) will underpin the performance and enforces us to rely on centralized server for computation, doing this will ruin the point of blockchain. To deal with this issue we came up with new architecture which eliminates the purpose of delete operation by maintaining two sets which are so called UTXO, and TXO sets.

Proposed architecture

We showed significant improved with existing state-of-the-art architectures, with a throughput of 165 transaction per second. Our architecture has a theoritical complexity for commiment update, transaction proof generation, verification, size update of \(\mathcal{O}(m)\), \(\mathcal{O}(k+d)\), \(\mathcal{O}(1)\), \(\mathcal{O}(1)\), and \(\mathcal{O}(k+d)\).

[LEFT] Performance of commitment updates [RIGHT] Perforance of commitment verification
Performance of propagation latency of a block in the network