October 6, 2026
The new, optional deduplicating media storage engine stores each binary or image file only once, using SHA-512-based deduplication. Content can stay in the database or be moved to S3-compatible object storage.
New deduplicating media storage engine (DMSE) for binary content
OpenCms 22 introduces the Deduplicating Media Storage Engine (DMSE), a new optional storage layer for media files such as images, PDFs and Word documents. Deduplication significantly reduces the database size, especially in larger OpenCms installations. The DMSE stores each binary file only once per installation: identical content is kept as a single copy, identified by its SHA-512 hash. This applies to the same file in the online and offline project as well as to the same image uploaded to several folders.
The DMSE supports multiple storage backends. Binaries can remain in the OpenCms database or be moved to S3-compatible object storage. Storage backends can also be switched later.
The effect on the database footprint is substantial: in typical installations, deduplication alone reduces the database size by around 40%, and moving binaries to S3 cuts it by 80% or more.
For editors and users, nothing changes. All content continues to be managed in the OpenCms Workplace as before. The DMSE operates transparently in the background, determining only how and where binary data is physically stored.
The DMSE is only available in the Jakarta variant of OpenCms 22.
Gradual introduction
For new OpenCms 22 installations, the DMSE is enabled by default.
Existing installations can also be switched to the DMSE as part of an upgrade. In this case, the DMSE can initially be operated in compatibility mode, where binary and image files continue to be stored in the same way as before. The deduplicating functionality of the storage layer can then be activated later.
Database or S3
Using the DMSE does not require an external storage system. The database storage backend provides deduplication of media files without requiring any additional storage infrastructure.
For installations with large or rapidly growing media collections, S3-compatible object storage can be used instead. This makes it possible to separate large volumes of binary data from the OpenCms database and manage them more flexibly.
Security and access control
Using an external storage backend like S3 with the DMSE does not change the OpenCms security model. Media files remain part of the OpenCms VFS and continue to be protected by the OpenCms ACL permission system. The storage backend only determines where the binary data is physically stored; it does not make protected content publicly accessible.
As with any external storage service, the storage backend itself must be configured securely.
Further information
With the DMSE, OpenCms provides the foundation for efficient and flexible storage of large media collections: fully transparent to editors and users, with no changes to the familiar way of working with the OpenCms VFS.
For more information about the DMSE, including configuration options, storage policies, available backends and the migration of existing content, please refer to the OpenCms documentation.