The necessity for backups has always existed, but the reason for backing up has changed significantly in recent years. Today, backing up data is just as important for cyber security reasons as it ever has been for disaster recovery. But our architecture must be rethought with this new emphasis.
When did we start conducting data backups?
A long time ago–in a galaxy far, far away…–backups we’re theoretically designed to mitigate against the risk of a disaster: fire, flooding, equipment failure etc. In reality, they were used primarily to correct bad decisions (we updated the server and it crashed, now we must go back to the previous version). A long standing practice of any IT change process I have been a part of has been “Back it up before you do that.” With the prevalence of virtual machines and the ease of taking a “snapshot,” back ups became very easy to do. Software and converged infrastructure have also made this increasingly robust and convenient as well.
However, with convenience comes a price. Many of our backup systems are on shared storage. We back up to the same place logically that our files are stored. And this is the underlying fallacy in our new cyber security reality. Our backups used to go to tape and get stored off-site. A return to this complexity needs to occur.
Backup Best Practices
Backups need to be on a completely separate storage volume that is not accessible to anyone or any bot, except that backup software. The credentials need to have strict complexity and policy to prevent access. Traffic should only be initiated from the backup network to the backup target and no traffic allowed to be initiated from the client network. Additionally, this information needs to be taken offline with regularity, removing it from the network.
Here’s a scenario: Organization X is performing backups and test restores according to their risk management profile. Some info is backed up daily, some hourly. Everyone is happy with the results. Suddenly, ransomware attacks the network and begins encrypting any data that is exposed, including backup files on a shared drive. This renders the backups useless for recovery from this attack.
Finally, this needs to be an executive level discussion. If you were the CEO of an organization, you would immediately be informed if the network was “down.” Being operational and ensuring your employees are productive is the most important piece of information you can receive from your IT team. The second most important piece of information should be “the backup process didn’t work last night.” The amount of risk this puts you in, potentially having to replace work from an entire day or longer, should be a risk you are aware of and constantly guarding against.