W3C Video on the Web Workshop
Position Paper
December 12-13, 2007
San Jose, CA
Background
Web video and associated rich-media content (e.g.,
3D virtual worlds, MMOGs) have already made the proverbial leap
from the incubation/hobbyist stage to initial mainstream adoption
with millions of consumers, participants and content choices.
However, these dynamics notwithstanding, current adoption
rates lag more mature Web technologies (e.g., HTML-based text,
graphics, audio) by several orders of magnitude. Further, the
quality and reliability of the user experience associated with Web
video varies widely.
While numerous factors clearly contribute to this
heterogeneity and varied adoption pacing, including relatively
recent advances in broadband availability and quality, efficient
video codecs, and commoditization and simultaneous performance
gains in consumer processing (CPU and GPU) and storage
technologies, a number of vexing challenges have arisen which are
unique to the deployment of high-bandwidth, real-time services such
as streaming Web video, interactive multiplayer games and virtual
worlds. The W3C’s new Web Video Initiative appears to be
positioned optimally to identify and address a number of these
issues, and this paper is offered to highlight several suggested
work items.
Problem Statement
Given the unique characteristics of Web video
services – namely, (1) bandwidth-intensive transport/distribution,
(2) CPU/GPU-intensive encoding/decoding/rendering, (3) very large
file sizes, and (4) wide variability in policies associated with
distribution, consumption and sharing – numerous technical,
business and economic challenges have emerged, representing an
inhibiting force in opposition to accelerated adoption. While
some of these challenges should appropriately remain within the
competitive domain, allowing for service provider and product
differentiation, others represent strategic enablers which may
accelerate both producer and consumer adoption of the Web as a
rich-media distribution and consumption medium, simultaneously
improving the quality of content, the consistency of user
experience and the overall economic viability of this emerging
sector.
A preliminary list of key market challenges (and,
conversely, potential enablers) includes:
- Appreciable
cost of high-bandwidth content distribution
(typically several orders of magnitude larger than
textual, static-graphic or audio content from a time-to-consume
perspective)
- Access, regional-area and backbone
network capacity management and planning
challenges; new economic and operational models
will likely be necessary in order to motivate continued capital
investment, particularly in the face of high-bandwidth Web video
and P2P file-sharing services
- Lack of
quantified network performance and QoS metrics and
assurances
; compelling network neutrality arguments
notwithstanding, not all IP traffic exhibits the same distribution
requirements (e.g., bandwidth, latency, jitter and TTL) and some
measures should be introduced to address this natural,
shared-network heterogeneity
- Pronounced
variability in the consumption and availability of
(shared or potentially shared) resources
(both network and consumer-based); from an
economic standpoint this represents inefficiency and waste in
current service models, while simultaneously offering the potential
for optimizations
- While underlying protocols (XML/RDF), frameworks
(REST/WSDL Web services) and tools exist, neither formal nor de
facto schemas or standards have emerged for capturing and
sharing
media metadata
throughout the distribution value chain
- Beyond the absence of a consensus media metadata
standard, no formal standard exists for consistently
identifying and naming media assets
; in the case of professionally-produced,
feature-length films as many as 50 versions of a particular asset
may be released, thereby severely compounding this problem
- While numerous, proprietary
digital rights management and conditional access
systems
have been widely deployed, in many cases these
systems have been used to discourage or prevent subscriber and
content mobility; a common, open-standard media policy
specification mechanism may help to mitigate this issue
- Content piracy and copyright infringement
continues to limit the availability of
high-quality content; automated mechanisms should be researched and
developed to address this concern
Proposed Work Items
Based upon these market considerations, the
following technical work items may prove compelling and valuable in
the context of the W3C’s current Web video incubation efforts.
It should also be noted that these further align with the
W3C’s overarching “one, fully-actualized Web” charter:
- Peer-to-Peer (P2P) architecture framework and
protocol design, emphasizing explicit network feedback, resource
negotiation and performance heuristics
- Ideally, in
collaboration with the DCIA’s P4P working group and the IETF’s P2P
efforts
- Explicit resource
management mechanisms (e.g., PCMM) should also be
considered
- Virtual economic models facilitating fare,
efficient and transparent allocation of network bandwidth and
client processing and storage resources
- Possibly, in collaboration with academic research
underway in this area at Harvard (SEAS) and elsewhere
- Media metadata schema standardization with
extensibility to Semantic Web technologies
- Ideally, aligned and in
ongoing collaboration with similar efforts at CableLabs, MovieLabs
and other industry standards organizations
- Digital rights policy specification
mechanism/protocol allowing content producers and owners to
unambiguously articulate fair-use rights and restrictions in a
platform-independent manner
- Ideally, secured but decoupled from underlying and
supporting usage and copy enforcement mechanisms
- Substantial research and innovation has been
underway in the areas of digital
watermarking and fingerprinting
to further mitigate inappropriate distribution and
consumption practices
CableLabs and its member MSOs have a significant
amount of interest in these and related topics, and I personally
look forward to participating in this collaborative industry
effort.
Jason Gaedtke
Chief Scientist
CableLabs