Abstract
Near-live P2P video streaming is already popular among
consumers, but typically this type of application is bound
to fixed Internet access, as streaming requires a sustained
bitrate over time and is less tolerant to environments
with volatile throughput rates. For users moving at speeds
larger than walking speed, P2P video streaming is hardly
possible, as the achievable throughput for a single peer
is insufficient to obtain the video. When multiple peers
are interested in the same video stream and they are
moving along the same path, they could cooperatively
retrieve chunks, instead of acting independently. The
participating peers will need to bundle their achievable
throughput and coordinate the chunk retrieval. We present
such a cooperative P2P video streaming system, introducing
a new scheduler and develop a light-weight throughput
estimator. 3G cellular link measurements and a simulation
study demonstrate the feasibility of our approach for a
train scenario.
Users
Please
log in to take part in the discussion (add own reviews or comments).