The rising demand for VoD and OTT services has made it imperative for content owners to protect their assets against leaks and illegal usage. While producers and distributors earlier promoted DRM protected content for maintaining security, digital watermarking techniques have now gained wide scale acceptance as an additional tool to deter illegal streaming and minimize damage as soon as possible. Watermarking helps create a uniquely identifiable stream for each viewing session. The approaches to watermarking can be broadly categorized into bitstream modification, A/B watermarking, and client-side watermarking.
Session-based watermarking can be performed to deliver unique content copies to each individual user on any kind of streaming player or device. Since all streaming sessions are delivered from the CDN edge, video watermarking functions can be run on the edge server. Service providers use a new two-step watermarking technology where a watermark can directly be embedded onto the encoded bitstream, while the edge server uses a unique identifier to deliver content copies to the users.
The key solution elements of anti-piracy measures when using an edge server are as follows:
- Watermark pre-processing: Adaptive bit-rate (ABR) transcoding is first carried out on the video asset to be delivered. This helps in inspecting the video and identifying locations in the video stream where changes can be made to encode a WMID (Watermark Identifier). The data obtained from this analysis is encoded as metadata to be encapsulated along with the bitstream during packaging.
- Delivery of the watermark token: The user’s video player requests for a WMID before any viewing session. A WMID is served as a JSON web token created by the content provider’s platform management subsystem.
- Embedding of watermark ID: When a video request is raised, the edge server verifies the authenticity of the WMID token. Once the compliance tests return positive results, the watermark embedding function is applied to the requested ABR video using the WMID given in the token. The CDN then releases pre-watermarked versions of the video to deliver a uniquely watermarked video stream to each client.
This bitstream watermarking technology using edge computing requires only a single version of the asset and thus reduces storage and bandwidth requirements, thereby decreasing video latency. In the event of leakage, the watermark can be identified and extracted to track the source of leakage and take necessary remedial actions.