I. Introduction 1 1.1 Introduction 1 1.2 Main contributions 2 1.3 Structure of thesis 3 II. Background 4 2.1 HyperLedger Fabric (HLF) 4 2.2 Resource separation of FastFabric 7 2.3 Re-simulation technique of XOX Fabric 7 2.4 Reordering technique of Fabric++ 11 III. Reducing Version Mismatch 13 3.1 System architecture 13 3.2 Necessity of state synchronization among peers 13 3.3 Adaptive state synchronization and early re-simulation 15 IV. Implementation 17 4.1 Adaptive state synchronization 17 4.2 Early re-simulation in endorsement peers 17 4.3 Chaincode embedding 18 V. Experimantal Evaluation 22 5.1 Experimental Setup 22 5.2 Successful transaction ratio 23 5.3 Throughput and latency 25 5.4 Impact of key ideas on performance 30 5.5 Scalability 33 VI. Related Work 34 VII. Conclusions 36 References 37 요약문 41