SAPUTRO, Edy and Mustafid, Mustafid and Suseno, Jatmiko Endro (2025) ALGORITMA KONSENSUS PROOF-OF-AUTHORITY BLOCKCHAIN UNTUK SINKRONISASI DATA KLASTER PADA KOMPUTASI AWAN TERDISTRIBUSI. Masters thesis, UNIVERSITAS DIPONEGORO.

	Text Cover-1.pdf Download (84kB)
	Text Cover.pdf Restricted to Repository staff only Download (2MB)
	Text BAB I.pdf Download (278kB)
	Text BAB II.pdf Download (971kB)
	Text BAB III.pdf Restricted to Repository staff only Download (1MB)
	Text BAB IV.pdf Restricted to Repository staff only Download (811kB)
	Text BAB V.pdf Restricted to Repository staff only Download (274kB)
	Text Daftar Pustaka.pdf Download (260kB)
	Text Lampiran.pdf Restricted to Repository staff only Download (1MB)

Abstract

Sinkronisasi data dalam sistem terdistribusi menghadapi tantangan dalam efisiensi, konsistensi metadata, dan konsumsi bandwidth. Penelitian ini mengembangkan framework Clustered Data Synchronization Framework (CDSF) dengan mengintegrasikan algoritma klasterisasi K-Means, metode pembobotan TF-IDF, blockchain berbasis konsensus Proof-of-Authority (PoA) dengan protokol Quorum Byzantine Fault Tolerance (QBFT), serta penyimpanan terdistribusi InterPlanetary File System (IPFS). Dataset produk e-commerce dari Kaggle digunakan sebagai objek pengujian untuk mensimulasikan skenario dunia nyata. Hasil klasterisasi diuji dengan metode Elbow dan evaluasi Silhouette Score untuk memastikan pemisahan antar klaster yang optimal. Metadata hasil klasterisasi disimpan di IPFS dan diregistrasikan ke blockchain menggunakan kontrak pintar. Implementasi mekanisme sinkronisasi selektif berbasis Content Identifier (CID) memungkinkan node menarik file hanya jika metadata telah berubah, sehingga mengurangi redundansi transfer data dan menghemat bandwidth. Pengujian menunjukkan rata-rata waktu finalisasi transaksi metadata pada QBFT berkisar 2–3 detik, dan efisiensi sinkronisasi mencapai penghematan bandwidth hingga 30%. Framework CDSF terbukti handal dan fault-tolerant dalam sinkronisasi data antar node dalam lingkungan komputasi awan terdistribusi.
Kata Kunci: Sinkronisasi Data, K-Means, TF-IDF, Blockchain, Proof-of-Authority, QBFT, IPFS, Komputasi Awan Terdistribusi

Data synchronization in distributed systems faces challenges in efficiency, metadata consistency, and bandwidth consumption. This research develops a Clustered Data Synchronization Framework (CDSF) by integrating K-Means clustering, TF-IDF weighting, Proof-of-Authority (PoA) blockchain consensus with the Quorum Byzantine Fault Tolerance (QBFT) protocol, and InterPlanetary File System (IPFS) for decentralized storage. An e-commerce product dataset from Kaggle is used to simulate real-world scenarios. Clustering results are evaluated using the Elbow method and Silhouette Score to ensure optimal cluster separation. Metadata from clustering is stored in IPFS and registered on the blockchain via smart contracts. The selective synchronization mechanism, based on Content Identifiers (CIDs), allows nodes to fetch files only when metadata changes, reducing data transfer redundancy and saving bandwidth. Experiments show an average metadata transaction finalization time of 2–3 seconds on QBFT, with bandwidth savings reaching up to 30%. The CDSF framework demonstrates reliability and fault tolerance in synchronizing data across nodes in distributed cloud environments.
Keywords: Data Synchronization, K-Means, TF-IDF, Blockchain, Proof-of-Authority, QBFT, IPFS, Distributed Cloud Computing

Actions (login required)

View Item