Big data are transferred, routed and forward to their destinations as a combination of a series of packets one by one, without considering the data in their entity in today's packet network. In this paper, we firstly summarize those transportation methods of bid data on packet networks and optical networks, respectively, and then propose a new data transportation architecture to collect, store and reshape the data flows in large time scale until it find the connection segments available on schedule. It leverages both circuit and storage merits together to form a now network node, which is termed as SSS in this paper. The intermediate storages act to relay and bridge these sub-connections into an equivalent entirety.