Another significant change in Oracle IRM 11g is that we now use XML to create the tamperproof header for each sealed document. This article explains what this means, and what benefit it offers.

So, every sealed file has a metadata header that contains information about the document - its classification, its format, the user who sealed it, the name and URL of the IRM Server, and much more.

The IRM Desktop and other IRM applications use this information to formulate the request for rights, as well as to enhance the user experience by exposing some of the metadata in the user interface. For example, in Windows explorer you can see some metadata exposed as properties of a sealed file and in the mouse-over tooltip.

The following image shows 10g and 11g metadata side by side.

As you can see, the 11g metadata is written as XML as opposed to the simple delimited text format used in 10g.

So why does this matter?

The key benefit of using XML is that it creates the opportunity for sealing applications to use custom metadata. This in turn creates the opportunity for custom classification models to be defined and enforced.

Out of the box, the solution uses the context classification model, in which two particular pieces of metadata form the basis of rights evaluation - the context name and the document's item code. But a custom sealing application could use some other model entirely, enabling rights decisions to be evaluated on some other basis.

The integration with Oracle Beehive is a great example of this. When a user adds a document to a Beehive workspace, that document can be automatically sealed with metadata that represents the Beehive security model rather than the context model. As a consequence, IRM can enforce the Beehive security model precisely and all rights configuration can actually be managed through the Beehive UI rather than the IRM UI. In this scenario, IRM simply supports the Beehive application, seamlessly extending Beehive security to all copies of workspace documents without any additional administration.

Finally, I mentioned that the metadata header is tamperproof. This is obviously to stop a rogue user modifying the metadata with a view to gaining unauthorised access - reclassifying a board document to a less sensitive classifcation, for example. To prevent this, the header is digitally signed and can only be manipulated by a suitably authorised sealing application.