Under what threat scenarios does this actually improve security? If the raw data is accessible to one process, it’s accessible to root, same as a normal db. Is the idea that only derived values can be queried once the data is in, and therefore there is no process with access to the raw data?
Under what threat scenarios does this actually improve security? If the raw data is accessible to one process, it’s accessible to root, same as a normal db. Is the idea that only derived values can be queried once the data is in, and therefore there is no process with access to the raw data?
“privileged attackers able to access a database’s memory […] for example, malicious administrators or rootkits”
Doesn’t sound like the most important threat to databases…?