In recent years, the high maintenance cost of centralized video on demand sys
tems has led to the development of peer to peer video on demand systems.
These peer to peer systems help to remove the cost and bandwidth limitations
of a centralized group of servers. In a peer to peer scenario, the publisher and a
small set of peers who were published to must handle all video requests. If many
peers request a video after it is released, the small set of peers with the video
cached may not have enough bandwidth to satisfy all requests. This situation
is known as a flash crowd. We propose a hybrid peer to peer framework that
allows publishers to publish videos before release time. For marketing purposes,
it is common for videos that are ready for distribution to be kept from being
released until a preset release time. By distributing a video before the release
time, more peers will have the video at release time, thus allowing more requests
to be handled. A hybrid peer to peer encryption management system is used
to prevent users from viewing videos before release time. In order to determine
who to distribute a video to before users are allowed to view the video, we design
a hybrid peer to peer subscription system. In this system, users may specify
interest in sets of videos and are notified of new videos matching the interest
so that retrieval may start. Finally, we modify an existing peer to peer video
on demand framework to better handle concurrent streaming and downloading.
Our experiments show that this framework can greatly increase a peer to peer
streaming system’s ability to handle flash crowd situations.
| Identifer | oai:union.ndltd.org:LACETR/oai:collectionscanada.gc.ca:BVAU.2429/4067 |
| Date | 11 1900 |
| Creators | Chiu, Stanley Kai Him |
| Publisher | University of British Columbia |
| Source Sets | Library and Archives Canada ETDs Repository / Centre d'archives des thèses électroniques de Bibliothèque et Archives Canada |
| Language | English |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
Page generated in 0.002 seconds