NAB
SMPTE, MovieLabs, and EBU have published a guide to working with cloud technology as part of the overall industry effort to standardize production and distribution in the cloud.
article here
The ‘Media in the Cloud: Ontology and Semantic Web
Technology Navigation Guide’ is
available for free on the SMPTE website. It is described as a primer on the
use of ontologies and other semantic web technologies characterized by the
movement of workflows into the cloud.
World Wide Web co-founder Tim Berners-Lee first proposed the
semantic web in 2001. This was the concept of moving from the mere presentation
of content on the web to actionable human and machine-readable data.
Movement towards such a web has been slow, details the SMPTE
guide, but may now be regarded as accelerating thanks to a convergence of
mechanisms and specifications that make it more practical and more desirable.
These include the cloud itself, which is not a requirement for the use of
semantic web technologies but is certainly a catalyst; microservices and application
programming interfaces (APIs); artificial intelligence (AI) and machine
learning (ML); and media-specific ontology specifications.
As the Guide outlines, the movement into the cloud comes
with attendant expectations around automation, agility, and scalability and
challenges in areas such as interoperability, portability, discovery, and
orchestration.
“It is an ever more data-driven eco-system with an increased
need for consistent, interoperable metadata and semantics to drive and manage
distributed processes and workflows. Fortunately, semantic web technologies —
which may be regarded as internet-native — have great potential to address some
of these challenges, supporting the combination of standards-based
machine-readability with the representation of human subject matter expertise.”
Unfortunately, there
is a lack of awareness about what semantic technologies are, what ontologies
are, where they are applicable, and how to deploy them.
“The shift of media workflows to the cloud — an ever more
data-driven ecosystem — yields many benefits, including greater automation,
agility, and scalability. But to realize these, organizations must successfully
address challenges related to workflow interoperability, data portability, and
the management of complex sets of assets,” said MovieLabs CTO Jim Helman.
“Media ontologies provide the essential knowledge framework to address those
challenges.”
What is an Ontology?
For the purposes of this paper, an ontology is a formal
model that represents a given “knowledge domain” — meaning the entities that
are meaningful within that space and the relationships between them — using a
set of specifications developed by the World Wide Web Consortium (W3C):
primarily RDF; RDFS; OWL; SKOS and SPARQL.
“In practical terms an ontology represents explicit business
knowledge and provides the scaffolding for the core data infrastructure of an
enterprise or a broader field,” says the Guide.
It goes on: Understanding and managing the complex
relationships between all elements in the content life cycle — from scripts to
assets to the tasks being performed across media workflows — require richer
metadata. An ontology provides a framework needed to support application and
service integration, asset and content management, and search and discovery,
among other functions.
The utilization of RDF in particular provides a mechanism
for information exchange between applications without a loss of meaning and for
creating linked data, meaning data that is interlinked with, and enriched by,
data from heterogenous and distributed sources.
The paper warns that Semantic technology is not a magic
bullet. It does not necessarily replace other technologies and is not a good
fit for every use case. It is not as mature as the relational database/SQL
eco-system, which may be more appropriate for predictable and consistent data
that is not characterized by many or complex relations, and for which there is
a greater pool of expert human resources.
However, we learn that semantic technology can be smartly
stacked with other tools and is in broad terms a good fit for cloud-based
eco-systems, having its genesis as a web-based approach to knowledge
management.
“Where there is likely room for development in the near
future is in the media factory — the production and distribution supply chain
and in media asset management and media processing workflows.
“As these migrate to the cloud or hybrid cloud/on-premises
and microservice-based deployments, it will become easier to integrate semantic
technologies into day-to-day operations, and more vendors will integrate them
into their offerings.”
The paper concludes with a thumbnail view of all the various
technologies discussed in this paper from APIs to Wikidata via Data Lake and
Node.
“Consistent and interoperable metadata and semantics are key for connecting data sets along the value chain, managing distributed workflows, and integrating applications. They are also crucial for content management and search and discovery,” said Hans Hoffmann, head of Media Fundamentals and Production Technology at EBU Technology and Innovation. “This navigation guide is the result of a great collaboration between EBU, SMPTE, and MovieLabs, three key actors in this field, and it will greatly help the media industry in its transformation into a data-driven ecosystem.”
No comments:
Post a Comment