I don't see any significant advantage in using Hard UART except for increased cost on hardware and size of hardware. Whats the harm in handling one device at a time and putting others on hold using handshake

Amazingly enough you should be able (with some fancy tapdancing) to handle several UARTs when bit-banging. The HW UART clocks at 16* bitrate and if your timer does the same (and the baudrate is low enough and the same for all channels) you should be able to sample several softARTs in each timeslice. Note, while possible, this will require "fancy tapdancing"

Please note I abhor bitbanged UARTs, however since the scheme is processing UARTs only it might be viable.

Erik