Before the Internet was common and cheap enough here in NZ, I was doing TCP/IP (very slow but still effective, 1200b/s half-duplex) by Ham Radio. On 386 and 486 machines (even 286's?) running DOS 5-ish, JNOS gave a semblance of linux's functionality and language (Slackware Linux could be up to 30x 1.4MB floppy-disks!, and required 386 or better processor), and with radio's fragile connections an occasional 'tcp kick' could refresh a connection. I'd venture a similar command could still be useful: 0.1 kB/s connections timing-out...
... or is this OS-level stuff? The internet's data transport is pretty much all Linux ...
Failing that, could there be a 'refresh' of search per file? a re-search without disconnecting existing connections? I've found some sources with long queues will allow me to re-join closer to the 'head' if I 'Stop' the file and 'Resume' - others re-join (if at all) at the tail
If a parallel search turns up shorter queues one could re-join there.
What think y'all?
F Bt