Stateless Blockchains for Scalable Transaction Processing

Blockchains enable mutually distrusting parties to maintain a common transaction ledger without a central authority, thereby empowering a wide range of innovative applications in healthcare, fintech, e-notarization, IoT, supply chains, and many other fields. However, most blockchain systems today require participant nodes to maintain the entire state of the ledger, which poses a heavy storage burden and limits scalability.

This project aims to explore a new paradigm of stateless blockchains for scalable transaction processing. The main idea is to store only short commitments of the ledger state on the chain, with the remaining (stateful) data being offloaded to off-chain storage nodes. The benefits are threefold. First, this approach can ease the storage burden of the blockchain network, making it more affordable to ordinary nodes. Second, it can reduce the bootstrap time of new nodes joining the blockchain network. Third, with the ledger state stored off the chain, costly transaction executions are shifted to storage nodes. Thus, blockchain nodes only need to validate the executions, and that incurs a much lower cost. Moreover, transactions can be executed on storage nodes in parallel, to improve system throughput.

There are several challenges to a realization of this paradigm. First, enabling the blockchain to validate off-chain transaction executions that may contain arbitrary logic is a fundamental issue. Second, conflicts can occur among the transactions, thus entailing new techniques that can identify the conflicts and maintain serializable on-chain commitments. Third, synchronizing the states among the storage nodes is another challenge. To address these challenges, our research agenda includes: i) design of a new block structure and a set of commitment protocols that support low-cost transaction validation in stateless settings; ii) development of efficient transaction processing algorithms that are capable of conflict detection and commitment serialization; iii) investigation of distributed network protocols for storage nodes to discover and merge new state updates; and iv) exploration of optimization techniques, such as off-chain storage sharding and recursive block attestation, to further improve system scalability. Finally, we will develop a proof-of-concept prototype system to demonstrate the performance of our proposals.

With our rich research experience in blockchain data management and verifiable computing, we expect the outcome of this project to accelerate the growth and adoption of blockchain technologies and decentralized services in the pertinent industries.