Proposal for GGIE Glass to Glass session at TPAC

My apologies for sending this today, I meant to send this last Friday, but I was off the grid while traveling home from Europe.


At TPAC, I would like to get 60 minutes on the IG agenda to present and discuss the ideas that the Glass to Glass Ecosystem - GGIE - group has been discussing.  As people may recall I gave a short presentation at the Web & TV IG Workshop in Munich. Following TPAC, I would like to continue the GGIE work as an IG project.

The essence of GGIE is how to do more and better content on the Internet through methods such as enabling devices to identify the content they are creating or viewing and to work with services and devices to make smart decisions on managing and accessing the content.

To enable smart content decisions devices/browsers need to be able to identify the content they are working with. This can be done by extracting information from the metadata in the container, but often this information isn't present so fallback methods such as extracting a watermark that was previously put into the content, or computing a fingerprint and comparing that against a fingerprint database are needed.   All these methods ultimately return an identifier that is unique to the content and that enable lookups for additional data on the content, or interaction with services that can provide services like caching, storage etc.

Once you've established a content identification scheme, application and device architects can innovate using this as a fundamental base to build upon..

There is work already going on in other standards groups around these ideas, and there will be need for coordination with groups like the IETF on standardized URNs which include the unique content identifiers. There is also watermarking/fingerprinting use cases already in discussion in the Web & TV IG that this compliments.

Some of the work will be enabling for services and protocols that would be done in other standards groups.

In keeping with the IG scope, this work will focus on developing use cases and identifying missing functionality.

The proposed items to discuss

1. Content detection API in the browser, which would support methods including:
   A. Watermark extraction
   B. Fingerprint generation and searching
   C. Metadata extraction from media
   + All of which obtain a content identifier.

2. Standardized URN for digital content naming.   Coordinated with IETF , maybe others too

3. A discovery protocol for advertisement of availability of content by URN, and an API for applications to interact with it.

regards
Glenn Deen
----
Director Networking & Distribution Technology, NBCUniversal | glenn.deen@nbcuni.com | 1-818-777-8123