Interbank payment and settlement flows have changed substantially in the last decade. This paper applies social network analysis to settlement data from the Bank of Japan Financial Network System (BOJ-NET) to examine the structure of transactions in the interbank money market. We find that interbank payment flows have changed from a star-shaped network with money brokers mediating at the hub to a decentralized network with numerous other channels. We note that this decentralized network includes a core network composed of several financial subsectors, in which these core nodes serve as hubs for nodes in the peripheral sub-networks. This structure connects all nodes in the network within two to three steps of links. The network has a variegated structure, with some clusters of institutions on the periphery, and some institutions having strong links with the core and others having weak links. The structure of the network is a critical determinant of systemic risk, because the mechanism in which liquidity shocks are propagated to the entire interbank market, or likewise absorbed in the process of propagation, depends greatly on network topology. Shock simulation examines the propagation process using the settlement data.