On Tue, Nov 17, 2020 at 09:08:49PM +0000, Stephen Farrell wrote: > I guess a lot would have to go wrong for a sustained period > for FTP to save the day, but I could just about imagine it > happening. It's not quite movie-plot time but fracturing in > the root stores causing HTTPS to not work everywhere, plus > some SSH bugginess that affected most clients and broke > rsync/SSH might do it. And as we've seen this year, now > and then stuff does hit the fan. I agree with you that this sort of cascade-of-failures event seems unlikely. But I've learned a few hard lessons in the last couple of decades and one of those is that removing our tried-and-true old bridges because new ones make them putatively obsolete sometimes comes back to haunt us. [ A timely example has presented itself under unfortunate circumstances. At this moment, the Baltimore County (Maryland) public school system's entire IT infrastructure is down due to an apparent ransomware attack. They can't email students/parents about it...because they shut down the old system. They can't call students/parents...because they shut down the old system. They can't text students/parents...because they shut down the old system. They've had to resort to social networks (urgh) and local news media to get the word out. It would have cost them almost nothing to keep an old, separate email server with sufficient capacity and an appropriate contact database running on a disused desktop PC that was physically disconnected until needed. But they didn't. ] So if the incremental cost of keeping an old bridge around, even if it's very lightly used, is small enough...then I think it's a good idea to keep it. Consider it an insurance policy against a low-probability event. ---rsk
