The following terms are used in this section in the following meanings:
encryption . The process of making data unreadable using some secret (see 'key')
decryption . The opposite of encryption
signature . A sequence of binary codes that is calculated based on the original data and a key to secure the content from undetected change.
key . A secret, depending on the type can be symmetric or asymmetric.
Typical symmetric keys are passwords. Symmetric keys are by their nature more at risk of being compromised if the secret key is stolen. Symmetric keys are usually used to encrypt large amounts of data as they are very fast.
Asymmetric keys are more secure and their structure is more complicated. Asymmetric keys generally consist of a private and public key pair. The owner of the key can sign data to be verified by recipient with public key. A correspondent can encrypt the data using public key to be decrypted by private key owner. Asymmetric keys are used to sign data, or encrypt small amounts of data as they are slower to process than symmetric ones.
certificate . This is a structure of secure data, which identifies a certificate owner. This is similar to having a user-name and password but more secure. A 'certificate' can be associated with a document as an alias of the X.509 certificate. Certificates are issued by a third-party, not the owner or recipient in the verification path, namely the Certificate Authority (CA). The CA's function is to guarantee that a receiver can trust data accompanied by a particular certificate. Certificates will contain a public key, but never the private key. These must be distinguished from certificates that are exported together with their private key in PKCS#12 format, these are different things.
user account space . This is a data area where a user can store private data. Only ODBC sessions and web page processing code which runs on behalf of a certain SQL user account has access to this area. This is useful for caching private security related information such as keys or certificates.
key reference . A short hand for a key, sometime called a 'key name'. This is a string to uniquely identify a key in a user account space.
key instance . An entity representing a key in PL, this is a not the key per se. The key instance is used in encryption or decryption routines.
The following algorithms are supported:
|triple DES (http://www.w3.org/2001/04/xmlenc#tripledes-cbc)|
Canonicalization algorithms with comments:
The keys can be temporary or persistent keys. Temporary keys are loaded only in memory an will be lost when the server is restarted. These are usually used for making symmetric session keys. Persistent keys are kept on the file-system or in the database and will be loaded upon server startup. These are asymmetric keys, certificates that belong to the user.
The location where persistent keys are stored depends on the key reference. If the key reference is a 'file:' URI, then the API's assume file-system storage, otherwise they will be kept in the U_OPT column of the SYS_USERS table as a serialized string. The API functions are described below. Whether keys are stored on the file-system or within the database they will have an in-memory representation which is used in crypto functions. The memory cache of keys is contained in the user account space, hence no user can access or reference another user's keys. Furthermore if a user account is removed all associated keys will also be removed if they were stored in the database. If the keys were on the file-system only the in-memory cache will be deleted.