الفهرس 
CHAPTER 1 INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 122 |  |  |
| | | from | 122 | | |
Abstract
Concurrent multipath scheduling aims to realize the highest efficiency of the network
bandwidth utilization. However, concurrent multipath transfer confronts challenges
such as long transmission delay, receiver buffer blocking, packet loss, and poor quality
of service which significantly degrade the transmission performance in the network. To
overcome these challenges, the proposed framework for adaptive scheduling of the
concurrent multipath transfer is introduced. This proposed framework depends on the
estimation of the bandwidth, path load, packet loss rate, and path delay for next
transmissions. The estimation of the bandwidth, path load, packet loss rate, and path
delay is based on the transmission information details which are sent from the receiver
by the path map. Then, the path score based on the estimated bandwidth, path load,
packet loss rate, and path delay is introduced for the best path selection of the next
transmission. The simulation results show that the proposed framework improves
throughput and reduces retransmission time-out. A proposed framework uses the endto-end delay of paths to distribute the packets over asymmetric paths in ordering policy.
The proposed framework detects network congestion in assigned path and reduces the
congestion window. Furthermore, the proposed framework predicts receiver buffer
blocking and deactivates the highest delay path that causes this problem. The simulation
results show that the proposed framework achieves highest throughput, lower entropy
and lower average application end-to-end delay than the previous algorithms.