This research examines the combined effects of sharding mechanisms on blockchain systems scalability and security levels for transaction flow and decentralization optimization. The research demonstrates how blockchain performance benefits from network sharding because it splits the network into parallelizable shards to process transactions concurrently while decreasing computing resource demands on nodes. Probabilistic modeling of the network using binomial and hypergeometric distributions measures security threats from malicious node dominance. Additionally, the analysis demonstrates that distribution via verifiable random functions provides equal shard assignments, which reduces security threats. The simulation results demonstrated a 12.39-second transaction processing duration while showing node 00035 participating in mining block ‘9dd6‘ at timestamp 776. The assessment demonstrates that the sharding achieves decentralized operation together with increased performance via checkpointed cross-shard communication and dynamic shard reassignment controls, which need future refinement. The research delivers practical recommendations for upcoming blockchain systems such as the Ethereum 2.0 network, which presents scalable and secure decentralized application capabilities.