EMC Documentum can be used to store large numbers of rich media files for a variety of purposes. Beyond the usual content management capabilities that can be applied to pictures, videos, and audio content (assigning custom object types, metadata, permissions, etc.), content transformation capabilities can be applied to generate thumbnails from pictures and low-res videos from multi-gigabyte video files, to name just a couple possibilities. And multiple Documentum product client options (e.g., Webtop, DAMtop, Media WorkSpace, D2) exist to provide out-of-the-box capabilities for accessing this content.
A Documentum client is not always the most desirable option, though. There are per-seat licensing considerations that might pose a significant cost for large number of users. Opening a company’s network to external users to provide access to such a critical system could be a significant security concern. And architecting a Documentum solution to geographically distribute this content could conflict with the corporate desire for a single central repository.
One alternative approach might be to publish the assets with their metadata out of the Documentum repository for use by an external application that makes the content available to both internal and external users in a fashion that better meets these needs. EMC’s Interactive Delivery Services (IDS) product is well suited for this publishing task. It can publish files, metadata, and even renditions to an external file system and database. Additionally, its architecture supports multiple target publishing, offering a way to deal with needs for geographical distribution and/or replication.
For example, a typical customer scenario might be to leverage Documentum capabilities to author, manage, and process rich media assets inside of a repository, then use IDS to publish out all “appropriate” images with their metadata according to a defined business process. From there, a custom-built web application could give users the ability to search these published images based on the metadata in a relational database. The published files and database could be located on servers in the appropriate external-facing portion of a company’s network as long as the firewall configuration allows communication between the internal IDS Source and external IDS Target. These published files and database can also be replicated using many different approaches and technologies to provide geographical distribution.
The five steps to setup this publishing approach typically look like the following:
1. Install the IDS Source on your content server and IDS Target on any of the other servers where you need files to be published according to instructions in the IDS Installation Guide. If necessary, ensure that communication port(s) used by IDS are opened if a firewall is located between these servers. (This is typical if the application servers are located in a network DMZ but the content management systems are internally located.)
2. Create a Publish Cabinet/Folder: IDS publishes assets stored in a WCM Publish cabinet or folder (Web Cabinet type of “wcm_channel” or Web Cabinet Folder type of “wcm_channel_fld”). It’s feasible to create a single cabinet or folder to store all assets. However, this doesn’t typically fit a business process that also allows for unpublished assets.
For example, assets may require initial review and/or editing before they can be published. Likewise, assets that should no longer be published (e.g., they are “retired” or otherwise should no longer be available for use) may still need to be stored in the repository. One approach to accommodate both needs is to store all assets in one or more cabinets and/or folders, then link assets to or unlink them from separate WCM Publish cabinets/folders based on whether they are to be published or unpublished.
3. Configure IDS: The IDS Configuration settings are described in the IDS User Guide, but particular settings of note include setting the object type filters, the file format filters, and the publishing asset metadata. All of these can be set in the Advanced tab in the Content Delivery Configuration properties.
- Property Export Settings: Select “Export Properties” to publish asset metadata to a database. Two options that may be useful in some cases are those to “Include contentless properties” and “Include folder properties”. In the “Additional Properties” setting, add any custom metadata for custom types that are published, plus any appropriate default metadata that may be useful, such as object_name, title, subject, and keywords.
- Property Table Name: This value is used in the table names created when IDS publishes metadata and its use may be database-specific. For example, in a SQL Server 2005 setup, a Table Name value of “SCS_Propdb” results in IDS tables named: dbo.SCS_Propdb_m (for IDS publish information), dbo.SCS_Propdb_r (for repeating attribute values), and dbo.SCS_Propdb_s (for single attribute values).
- Content Selection Settings – Formats: One important use of file formats can be to include the system generated renditions for assets, such as JPEG thumbnails (which can be used as thumbnails by a custom website) and even XMP information extracted into XML renditions. Be sure to add all of the Documentum JPEG formats (e.g., jpeg_th, jpeg_preview, jpeg_lres) to ensure all renditions are published with the asset.
4. Run IDS Publish: From DA, run “End to End Test” to verify connectivity to the IDS Target(s). Once that works properly, perform a full publish that includes creation of the schema. (Check “Refresh entire site” and “Recreate property schema” boxes.) These options don’t need to be checked for subsequent publishes unless the schema changes (e.g., changes to the custom type metadata). Run IDS Configuration Publish manually while testing, but schedule the corresponding Documentum job automatically created for each IDS Configuration to begin automated publishing.
5. Publish Results: After a successful publish, the files that will appear in the IDS Target_Root folder depend on the object types and formats chosen to be published. For example, assets and renditions published from the repository are shown below:
In addition, metadata for those images appears in a database in the three standard IDS tables whose names are based on the IDS Configuration Property Table Name setting as previously described:
Note: This simplified three-table IDS structure works across different databases, but it may not provide the easiest way to query for information. In some cases, it may be easier to create database views on these tables that make it easier to perform specific searches, such as searching for assets of specific custom object types or search for those with specific metadata values. This approach may require extra database setup to ensure that views are automatically updated after an IDS publish is completed, though. For example, Oracle materialized views are one way to achieve this.
That’s it. IDS now helps implement the company’s business processes as the content distribution/delivery component that ensures assets are only published (and unpublished) when appropriate. It coordinates the flow of content between the Documentum repository where authoring takes place and any external solution built upon published files put onto an external file system and their metadata put into an external database. And now there are significantly more options for making that content available in ways that may better meet a customer’s needs than through the use of a Documentum client.