Encryption

PDF documents have built-in support for securing them by encrypting the content and assigning usage rights. The main access point to this facility in HexaPDF is HexaPDF::Document#encrypt.

How the Encryption Works

The PDF specification defines the algorithms that are used for encrypting, namely RC4 or AES, and how the encryption should be done.

RC4 is an old and nowadays insecure algorithm and should be avoided. AES is much better. Both algorithms can be used with various key length: RC4 from 40bit to 128bit and AES 128bit and 256bit. The latter, AES 256bit, was only standardized with PDF 2.0. Prior to this it was an Adobe extension. So for now the best option is AES 128bit to get good security with the broadest compatibility. This is the default setting for HexaPDF.

When a PDF is encrypted, all strings and byte streams (e.g. the page contents) are encrypted. Meta data streams in XMP format can be exempt from the encryption so that tools that parse the PDF file can still locate and extract them. This means that without the proper password viewing of the PDF is not possible.

The encryption key itself is provided by a so called security handler (see next section). This means that the “how”, i.e. how the encryption is done, is always the same, only the derivation of the encryption key is different.

Security Handlers

Security handlers provide the encryption key to the encryption facility and can specify additional information.

The standard security handler as defined by the PDF specification needs to be implemented by every PDF reader application and calculates the encryption key from a password. Actually, one can define two passwords:

In addition to deriving the encryption key from the passwords, the standard security handler provides a way to set usage rights, e.g. printing allowed, copying allowed, and so on. These usage rights should be enforced by a PDF reader application if the document was opened with the user password but not when opened with the owner password. See HexaPDF::Encryption::StandardSecurityHandler::Permissions for a list of all available permissions.

It is also possible to just assign usage rights. Then the PDF document is still encrypted but a user doesn’t need to enter a password for viewing the PDF document. However, the usage rights are still applied.

In addition to the standard security handler there is also a security handler for creating the encryption key using certificates. However, this functionality is probably only used in enterprise settings where a certain PDF reader application is mandated.