Date-dependent key-strategyThe application server needs to load keys temporarily into memory in order to work with the data. Otherwise it would not be able to perform efficient queries, for example. In order to prevent an "evil administrator" from obtaining access to all the data by simply dumping the application server's memory in this moment, Cumulus4j uses multiple keys (a few thousand). This way, the "evil administrator" could only gain access to a subset of the data (hopefully very little). One possible strategy to assign keys to data is by the time the data is written. At the moment, this is the only strategy supported by Cumulus4j. Other strategies might be supported later. The date-dependent key-strategy works by creating all keys for a certain period of time (by default 50 years) in advance. Every key is active (i.e. used for encryption) in a certain pre-defined time range (by default 1 day). Whenever data is encrypted, the key active at this moment is used to encrypt the record. The key-ID that was used to encrypt the data is then stored together with every record (so that the system knows which key it needs to decrypt the data later). Therefore, an "evil administrator" dumping the memory of an application server, will only be able to access a small subset of your data (e.g. usually a few days - in the worst case - e.g. if you just ran some report - maybe a few months). |
Documentation
AboutProject DocumentationBabelReleases
|
Deutsch