Jan 9, 2019 02:10 AM
The Data Storm (my term) arises when there is a system failure. Assume we can restore all data from backup to a new system in (say) x hours. In the best of all possible worlds, during those x hours we continue to acquire new data (x hours worth) which has to be restored. The very term 'system failure' suggests we won't be getting anything useful from it. Cry 'Havoc!.' and let slip the headless chickens of system failure. It may be that you can never catch up with the accumulating data. Or, more likely, someone makes a value judgement - we will restore from (whenever) to now and anything lost is - lost.
I had the experience of my server failing, within a few weeks my backup server failed and a few weeks later my backup backup server failed (all i386). Being a nerd I was playing with arm cpu's (Raspberry Pi's) and was just at the point when I could slot in a Raspberry Pi. Kind'a looked good but in reality it was pure luck.
The problem isn't just technical - imagine the accounts department of a bank being asked to pay for four backup systems - I don't think so. Ultimately a major bank will fail - quite possibly in an unrecoverable way. Your internet banking (savings) lost forever.
I had the experience of my server failing, within a few weeks my backup server failed and a few weeks later my backup backup server failed (all i386). Being a nerd I was playing with arm cpu's (Raspberry Pi's) and was just at the point when I could slot in a Raspberry Pi. Kind'a looked good but in reality it was pure luck.
The problem isn't just technical - imagine the accounts department of a bank being asked to pay for four backup systems - I don't think so. Ultimately a major bank will fail - quite possibly in an unrecoverable way. Your internet banking (savings) lost forever.