This article addresses the issues of scalability and security in blockchain networks, with a focus on sharding algorithms and decentralized data storage. Key challenges include the low throughput and high transaction latency in public networks such as Bitcoin and Ethereum. Sharding is examined as a method to enhance performance through data distribution, but it raises concerns regarding node management and reliability. Sharding schemes, such as Elastico, OmniLedger, Pyramid, RepChain, and SSchain, are analyzed, each presenting its own advantages and drawbacks. Alternative architectures like Directed Acyclic Graphs (DAGs) demonstrate potential for improved scalability but require further refinement to ensure decentralization and security. Protocols such as Brokerchain, Meepo, AHL, Benzene, and CycLedger offer unique approaches to addressing performance and transaction consistency issues. This article emphasizes the need for a comprehensive approach, including dynamic sharding, multi-level consensus, and inter-shard coordination. Additionally, a conceptual model is proposed that incorporates the sharding of transactions, states, and networks, which enables greater scalability and efficiency.