I don’t think data roaming is a solution to what is just simply bad or non-existent contigency plans.
When things go wrong in the middle of the night, I am quite prepared assign cause to human error. That’s when changes occur. And if that change knocks out the core backbone network, then big problems. Diagnostic data stops being sent to management servers unless on totally separate links. Management servers cannot communicate with network devices, unless totally separate links used.
And here’s the real problem. Support people cannot get to management servers because the network is down. You have to be onsite, and physically access devices.
And those who can be onsite, like first level network technicians may well have to consult higher levels, or access documents on file servers. Oops. No network, no phones.
Many big companies run their systems in ‘dark’ datacenters. There are no technical support people there at all overnight. Just a few security guards, and a list of numbers to call if an alarm comes up.
