الفهرس يوجد فقط 14 صفحة متاحة للعرض العام
 |  | from | 136 |  |  |
| | | from | 136 | | |
المستخلص
Due to the explosive growth of Internet communication and the increasing demand for lultimedia information on the web, delivering real-time video over the Internet becomes 1 important issue for many Internet multimedia applications. Transmission of real-time deo has bandwidth, delay, and loss requirements and the application-level quality for deo streaming relies on continuous playback, which means that neither buffer underflow ’r buffer overflow should occur. Since the current Best Effort (BE) network such as the temet does not provide any Quality of Service (QoS) guarantees to video transmission er the Internet, therefore, mapping the application-level QoS requirements into network¬reI requirements is required to achieve limited delay jitters and End-to-end application ’el QoS through adaptation. This imposes a need for temporal and spatial guarantees to iver the video streams over Internet Protocol (IP) networks to attain highest video Ismission rate, and lowest packet loss rate and end-to-end transmission delay from the ierlying network.
fhe distributed multimedia applications can suffer severe performance degradation er conditions of unpredictable delays and losses in the network. Consequently runtime ; adaptation is necessary to enable operation of these applications with acceptable ’ormance. During transmission of video streams over lP, video transmission rate may , due to load changes, packet loss disrupts the video coding, while transmission delays
decrease the throughput or cause underflow events of the receiving buffers for lding.
fenerally speaking, there are three main kinds of drawbacks arising during the imission of video streaming over the Internet:
ne packets arrive at destination in different orders, due to the fact that the bandwidth ilability is highly unpredictable because of lo@ ���<� he video streaming lals which make the quality adaptation difficult.
U’iable network latency, due to variable performance of the network caused by load Iges. More specifically, available bandwidth exhibits high variability. This can be t with, to some extent, using a buffer. However this option is limited by the buffer ~e available at the client.
e network congestion, due to low arrival rate and thus, the packets may be damaged 1St.
’IS thesis presents a new two adaptation schemes that enable multimedia applications form satisfactorily under constrained system and network resource availability.
~ first adaptation scheme is presented to analyze the effect of the TCP parameters on nsmission rates (higher and lower) than the default transmission rate to optimize the {ity of transmission. The adaptation scheme reaches to the optimal bandwidth that ~s the variation in the network latency due to the load changes to maintain the y for the network performance. The quality adaptation becomes easy by predict the lIe bandwidth using TCP window size which controls how much data can be in the k at anyone point. If it is too small, the sender will be idle at times and get poor
ance. By adjusting the TCP window size and buffer lengths to different values the interval times, it is possible to measure the bandwidth for every interval and e total bandwidth for number of streams.
e easiest way to determine the round trip time is to use ping from one host to another .~ the response times returned by ping. theory, the TCP window size is determined by the following equation:
rcp window size = the available B W of the network x round trip time
The first drawback results from highly unpredictable bandwidth available in the video ing traffic, which results from continuously load changes and therefore the ptation becomes difficult. According to this drawback, the packets arrive at their stination randomly. During the adaptation process, the thesis used the Jperf as a :timulation tool to measures MBW for the TCP protocol by allowing tuning of various ,parameters. Several options in packet can be configured in packet header with adjusting the : regular time interval to print information, easily. Using this scheme, the first drawback can P be overcome.
The second adaptation scheme is presented to analyze the effect of UDP parameters on delay jitter and datagram loss values to increase the efficiency of UDP protocol to prevent the network congestion and increase the adaptivity of the transmission. The adaptation scheme reaches to the lowest delay jitter and packet loss. By adjusting the UDP packet size and UDP buffer lengths to different values through the interval times, it is possible to measure the delay jitter and packet loss for every interval and also the total values for number of streams.
The second drawback results from the variable network performance due to load changes and as a result variable network latency will occur. More specifically available bandwidth exhibits high variability, which can be dealt with, to some extent, by using a buffer but that option is limited by the buffer space available at the client. The network congestion is the third drawback result from low arrival rate and also from the packets which damaged or lost. During the adaptation process, Jperf measures the delay jitters and datagram loss values by allowing the tuning of various parameters. from the results, the introduced system can predict the lowest values for the delay jitter and packet losses to prevent the network congestion. Using this scheme, the introduced system can overcome the second and the third drawbacks.
Using the mathematical equations, the optimal window size and the maximum acceptable delay jitter values can be calculated and compared with the experimental results.