Title :
Superchunk Based Fast Search in P2P-VoD System
Author :
Wang, Danqi ; Yeo, Chai Kiat
Author_Institution :
Centre for Multimedia & Network Technol., Nanyang Technol. Univ., Singapore, Singapore
Abstract :
It is crucial to provide reliable and fast search in P2P VoD system to enable user interactivity under peer dynamics. We propose a hybrid content discovery mechanism: superchunk based fast search network (SURFNet), based on our observations of current P2P VoD systems. SURFNet divides movies into superchunks and chunks, and takes advantages of stable peers to construct an AVL tree, which provides superchunk-level data availability information and aggregates superchunk holders into holder-chains. This structured overlay is further extended by a gossip-based unstructured network for chunk-level information exchange and data transmission. Our analysis and simulation results show that SURFNet supports efficient search and reduces seek latency significantly.
Keywords :
content-based retrieval; peer-to-peer computing; video on demand; video retrieval; P2P-VoD system; SURFNet; hybrid content discovery mechanism; peer dynamics; superchunk fast search; Aggregates; Algorithm design and analysis; Availability; Computer network reliability; Computer networks; Delay; Motion pictures; Multimedia systems; Peer to peer computing; Reliability engineering;
Conference_Titel :
Global Telecommunications Conference, 2009. GLOBECOM 2009. IEEE
Conference_Location :
Honolulu, HI
Print_ISBN :
978-1-4244-4148-8
DOI :
10.1109/GLOCOM.2009.5425514