We indicated notification as an interface, but this may be thought of as more of a “service” or feature than an API. Other features typically found in video search services include:
• Favorites - saved searches where results may appear on the search engine landing page freshly executed with each visit.
• History – access to videos recently viewed.
• User preferences - capture and support user preferences such as explicit content filtering. In some cases bandwidth or media player preferences are stored as a cookie.
• Cut-and-paste links - The service may make available markup suitable for rapid cut-and-paste into blogs or other applications to allow either link back to the video search engine with the video of interest selected, or to insert a video plug-in to deliver the video.
• Download - The ability for users to download the video with transcoding as appropriate for particular devices.
• Annotation - ability for users to post persistent comments, rate or tag videos and make these annotations searchable.
Conclusion
We’ve seen that the basic structure of video search involves acquisition, media processing and retrieval systems. This basic architecture has parallels in video on demand systems and media asset management (MAM) systems, albeit with different optimizations for the scale and features desired. We presented options for video search engine architectures and supported features and APIs along with the associated tradeoffs for system resource utilization. Architectures have evolved as new technologies have become practical and widespread and we will no doubt continue to see advances in video search architectures going forward.
