74595 Serial In/Parallel Out schematic

Output port using some old timey logic and the SPI port on a PIC.  On a PIC16F876 running at 20mhz, 5Mbps is the limit, on a 40mhz '252, you can pipe data out at 10Mbps.  There is no limit on the number of devices that can be chained together, all running off the same chip select.