A video server in which each media segment file sever (MSFS) stores media segment files (MSFs) forming video data, and sequence control brokers (SCBs) installed between the video server and subscribers terminals receive video requests from the subscriber terminals, and read out the requested MSFs from the MSFSs. The readout MSFs are exchanged by an ATM switch to be transmitted to the SCBs respectively, which outputs the MSFs at predetermined intervals to the request-sender subscriber terminals.