Alternative to scp, transferring files between linux machines by opening parallel connections
You could try using split(1) to break the file apart and then scp the pieces in parallel. The file could then be combined into a single file on the destination machine with 'cat'.
# on local hostsplit -b 1M large.file large.file. # split into 1MiB chunksfor f in large.file.*; do scp $f remote_host: & done# on remote hostcat large.file.* > large.file
Similar to Mike K's answer, check out https://code.google.com/p/scp-tsunami/ - it handles splitting the file, starting several scp processes to copy the parts and then joins them again...it can also copy to multiple hosts...
./scpTsunami.py -v -s -t 9 -b 10m -u dan bigfile.tar.gz /tmp -l remote.hostThat splits the file into 10MB chunks and copies them using 9 scp processes...