Knapsack problem based piece-picking algorithms for layered content in peer-to-peer networks
1. KNAPSACK PROBLEM-BASED PIECE-KNAPSACK PROBLEM-BASED PIECE-
PICKING ALGORITHMS FOR LAYEREDPICKING ALGORITHMS FOR LAYERED
CONTENT INCONTENT IN
PEER-TO-PEER NETWORKSPEER-TO-PEER NETWORKS
Michael Eberhard
Hermann Hellwagner
Christian Timmerer
Institute of Information Technology
Klagenfurt University
Austria
Michael Eberhard 1Knapsack Problem-based Piece-Picking Algorithms
This work is supported in the context of the P2P-Next project (FP7-ICT-216217)
Tibor Szkaliczki
László Szobonya
eLearning Department
MTA Sztaki
Hungary
2. PIECE-PICKING ALGORITHMSPIECE-PICKING ALGORITHMS
In a Bittorrent-based P2P system, the piece-picking
algorithm decides when to download which pieces
Bittorrent’s piece-picking algorithm (rarest-first) is not
suitable for video streaming or layered content
Require an algorithm that finds the best trade-off
between smooth playback and showing the best
possible quality
Michael Eberhard Knapsack Problem-based Piece-Picking Algorithms 2
3. KNAPSACK PROBLEM-BASED ALGORITHMSKNAPSACK PROBLEM-BASED ALGORITHMS
The piece-picking problem is closely related to the
Knapsack problem in combinatorial optimization
We analyzed existing algorithms for solving the
Knapsack problem and evaluated if they are applicable
for the piece-picking problem
We tried to improve the Knapsack Problem-based
algorithms considering specific P2P requirements
Michael Eberhard Knapsack Problem-based Piece-Picking Algorithms 3