PERFOMANCE OPTIMIZATION OF TRANSACTIONS PROCESSING ON BLOCKCHAIN TON

Semenyuk Timofey Andreevich

This article explores architectural and technical approaches to optimizing transaction processing performance in the TON blockchain ecosystem using a DEX protocol as a case study. The study addresses challenges such as high latency, gas fees, and the limitations of reactive trading under traditional interactions via external wallets. As a solution, the article proposes a transition to an intent-based interaction model, where users generate an intent to perform an operation without explicitly signing each transaction at the time of submission. This approach allows batching of requests, executing only essential steps on-chain, and significantly reducing network load. The core element of the proposed architecture is the SmartAccount, which enables non-custodial and programmable control over user funds. It operates alongside an embedded wallet and a sequencer system responsible for processing and finalizing user intents. The sequencer collects and orders intents in an off-chain execution queue, performs pre-validation, deduplication, and signature verification, and forms aggregated ExecuteBundles that are committed on-chain via a domain-specific smart contract. This results in high scalability, minimal response latency, and zero slippage. Comparative performance assessments are presented, along with recommendations for applying the intent-based processing model in high-load decentralized systems.

Keywords:blockchain, distributed ledger, transactional systems, distributed systems.