If you need to update a published session report or add attachments (images, files, links), send a note to osullivan@dvb.org.
How should DVB-I service discovery work at Internet scale
Host: William Cooper Reporter: William Cooper Session goal: Explore the market realities of Service Discovery using the DVB-I specification.
Summary of discussion:
We have a specification for DVB-I Service Discovery and Programme Metadata, but how do we implement Service Discovery internationally to support an open and competitive market ecosystem?
This session raised a number of open questions about practical service discovery that remain unresolved or intentionally beyond the scope of the DVB-I specification.
These issues may be solved without changing to specs but may require implementation recommendations.
Actors involved
- Viewers
- Manufacturers
- Legislators and regulators
- Broadcasters and media providers
We also consider bad actors – what could disrupt the integrity of the ecosystem or erode consumer confidence.
(See also session led by Raj on Bad Actor Journey)
Possible models
- Central – an entity or well-known address that offers a single point of service discovery
- National – broadcaster consortia led as currently being pursued in some markets
- International – multinational services such as those being offered by satellite operators
- Federated – scalable internet architectures tend to follow models of federation and delegation
- Manufacturer – potentially including their own online channel offerings
- Platform operator – independent aggregator or packager
- Personal – custom lists or individual services like a security camera
Any single central model presents problems of becoming a gatekeeper or arbiter and the responsibilities and liabilities that go with that.
The practical market reality is that clients may need to support a mixture of multiple models, possibly in different tiers. Plurality may be a precondition for a competitive market.
The internet tends to be international, distributed, and federated. We need to look at models that work, like DNS.
Any system needs to reflect regulatory requirements that have evolved through broadcast transmission and licensed spectrum and translate this into internet domain, which inherently behaves differently.
Service discovery needs to work not just for fixed televisions but for also for portable screens that may roam between locations.
The current focus appears to be on national markets but we may also need to consider international broadcasters, users with interests that span territories, and diaspora populations. How will fast channels work in this world?
Market trials have been organised by broadcasters looking to preserve existing business models and practices, but we need to consider the requirements of an open competitive market and consumer expectations.
We have the concept of a regulated list but it is meaningless without some authority or attribution to a relevant regulator, which may be national.
If other services are appended to a regulated service list its status becomes unreliable or misleading.
There may be a benefit in signalling that individual services are licensed by an identified regulator in a particular jurisdiction. This may require additions to the specification and out of band policies and processes.
We need to focus on user experience and expectations. Manufacturers and their products are key to this.
Next steps:
Any addition suggestions or questions can be added to the unconference wiki.
William to pull together a paper to go to CMI in April for further discussion to explore some ecosystem requirements and best practice deployment models that do not necessarily require any changes to the specification.