Performances Of Video Chat Applications Under Congestion Computer Science Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

For the performance of video chat applications let us analyze some reaction changes on Skype, Eyebeam, XLITE…etc. Usually we have a scenario where we have the speed of uplink is always lower than the speed of the downlink. Hence we have a difference of speeds during streaming as we not only receive videos but also we have to send them. Apart from these we also have a scenario where we have cross traffic which also causes congestion.

The adverse effect will be on the uplink can be justified from the above hence respective mechanisms should be followed for minimizing the congestion during the uplink transmission. The video and the audio qualities will have their own adverse effects when we have congestion during transmission. However there can be a comparison made between wireless loss and the congestion loss.

In the congestion loss we have loss of packets during transmission of bits both in uploading and also in downloading, but the major adverse effect can be cited during the uplink of transmission of bits. When this is compared to Skype the packet loss attained is relatively low as we have the usage of different bandwidths at different steps.

Related work of comparison between different sites:

Let us consider YouTube, Hulix. Buffering of the content is done before starting off and all the data is stored in bit rates and the most effective one from all is saved and utilized. Always the heterogeneous environments are selected for transmission. UDP suits to this kind of environment hence always transmission through UDP is preferred to TCP.

Another alternate method can be used for minimizing the occurrence of congestion where the transmission rate of bits is adjusted accordingly. Under utilization of this rate can also has an adverse effect like packet loss usually during uplink of data.

When there is a comparison made between the audio and the video applications occurrence of packet loss is seen maximum in video transmitting. We have two different losses here in bit transmitting which are the random losses and the congestion losses. Usually only packet loss doesn't refer to congestion occurrence.

There are many examples for the occurrence of congestion without any packet loss in wireless environment. In the wireless environment usually congestion occurs due to loss of connection, co-channel interference. Distinguishing between these losses is very much important as if there is any scenario cited where we have congestion but we don't know the reason for its occurrence.

Let us consider one example where we have congestion due to loss of connection but the user assumes this occurrence of congestion due to packet loss, thinking so he transmits the data at slower rate which finally results in poor quality of bit transmission.


For measuring these let us assume a desktop PC for adjusting the available bandwidth with two Ethernet cards and with all the set up fixed we have a dummy net application initiated. We have the PC's connected to the internet and we have all the traffic running through the dummy net. The dummy net is used as gateway.

Measuring with a Change in Bandwidth:

We have some special functions followed stepwise for measurement in here. Here we update with the bandwidth value in every step. Results are compared with Skype, Xlite and Eyebeam. The only advantage in following this is the description of the measurement is given in a more detailed form.

Let us consider behavior of Xlite after the measurement, it doesn't support this regular updation of bandwidth hence we have packet loss here. Even though we have a packet loss , the video streaming doesn't stop here and there will be no dropping of call here.

An eye beam uses H.264 codec same like an XLITE. When it faces congestion then eyebeam uses FEC technique such that audio and video runs satisfactorily. This technique can only support some kind of congestions but it cant support other type of congestion. If the other congestion occurs then eyebeam will maintain its same transmission no matter how the congestion would .One interesting aspect is that eyebeam receives more fluctuations compared to the XLITE .It will not lower its transmission and not allow other cross traffic to occupy it bandwidth.This would cause a heavy packet loss.

Whereas, the Skype will increase or decrease its transmission rate according to the bandwidth allocated. When the congestion occurs it will decrease its transmission rate according to the bandwidth allocated.

If the congestion occurred in the windows live messenger then it lowers its transmission rate until the congestion is resolved . Once the congestion is resolved it increases its transmission rate and get back to the original state but it takes long time to reach its position.

These four video chat applications react differently when the congestion occurs as Skype and Windows live messenger will lower its transmission rate and increases back when the congestion is done. Whereas, the Eyebeam and XLITE will not lower transmission rate and faces a heavy packet loss. So, it is proved that Skype and the live messenger are the video chat application which gives the user satisfaction as maintain the audio and video streaming good .And whereas the eyebeam and XLITE are not the good video chat applications as they faces a heavy packet loss.

A video chat application is said to be good when they keep their codec as low as possible and audio should be good. These video codec's should need to adapt the changes according to any requested bit rate.


Let us analyze how these different video chat applications respond when there is HTTP as cross traffic.

XLITE and eyebeam will not lower its transmission regardless of the cross traffic and faces a heavy packet loss. Skype reacts differently by lowering its transmission rate in accordance with the bandwidth allocated. Once the cross traffic finishes, Skype will retransmit its bandwidth quickly and reaches its original position. Windows Live Messenger will also lower its transmission rates according to the bandwidth allocated and cross traffic as an HTTP. Once the cross traffic finishes it retransmit to its original position but it takes long time to reach the position.


Let us analyze the performance of the different video chat applications when bit torrent acts as the cross traffic.

Eyebeam and XLITE will act as the same way in HTTP as it will not lowers its transmission whether or not the cross traffic is present. But the fluctuations in eyebeam will be more compared to the XLITE. One of the interesting aspect is that eyebeam and XLITE will not lower their transmission and also they will not allow bit torrent to allocate the bandwidth.

Skype will lower its transmission as soon as the bit torrents start. This limit will make the skype to lose audio and video streaming. As soon as the bit torrent finishes it will retransmit again and get back to its original position quickly. The reason why the Skype lowers its transmission because bit torrent occupies about 85% of the TCP connections.

Windows live messenger will act differently when bit torrent as cross traffic. It lowers its transmission and faces a packet loss. It re transmit the bandwidth as soon as the bit torrent finishes and regains its original position. It takes more time to regain its original position.

Different losses

Always when a packet loss occurs it will degrade the whole quality of the video chat. This is would be true when codec like H.264 is used with the correlation of frames. So, to resolve this loss a bandwidth adaptation algorithm needs to be implemented such that transmission rate is lowered to avoid the congestion. But not all the losses are due to the congestion as because there can be loss with the wireless networks. Different techniques like FEC needs to be implemented. But these techniques cannot completely resolve the problem. So, the differentiation needs to be done with the random loss and packet loss.

Let us try to analyze as how the different video chat applications can determine these losses by creating a dummy net which has 1% of random packet loss with two scenarios. First scenario is to introduce loss at start of the video chat session and other is to introduce at the middle of the video chat session.

Both Eyebeam and XLITE do not change their transmission rates. Skype reacts differently to these scenarios. When it is introduced in the middle of the session it reacts by FEC technique and when the losses is of congestion then it lowers it transmission rate. Skype can differentiate between random loss and congestion loss by monitoring the packet delay. If the delay is of congestion then it lowers its transmission rate and where as the random loss the transmission rate will be hiked by 20%

Future work

By looking at the performance of the different video chat applications I would like to improve the video chat session by introducing new techniques such that the streaming doesn't stop even when the congestion occurs. A new algorithm needs to be introduced such that allocations of bandwidth should go by the priority. Using this algorithm the bandwidth is allocated more to the video chat session such that video chat session will not have fluctuations. Priority will also be given to the audio and video such that it should loss intolerant and loss tolerant.


Until now we have discussed how the Skype, Windows live messenger, Eyebeam, XLITE video chat applications perform under cross traffic ,congestion, different losses and how they react in their transmission rates.

We came to know that skype changes its transmission rate according to its bandwidth either by increasing or decreasing. Skype also monitor different techniques RTT and jitter .It can also determine the difference between congestion loss and the random loss. It can also adapt the changes in the transmission rates according to the bandwidths.

Windows Live Messenger will lowers its transmission rate when the packet delay is occurred and re transmits to its original position. But it takes more time to retransmit to its original position.

XLITE and Eye beam will not lower the transmission rate when there is cross traffic. But interesting thing when the bit torrent acts as the cross traffic Eyebeam will be receiving more fluctuations compared to the XLITE .And one more interesting thing is that XLITE and eyebeam will not lower its transmission and also will not allow bit torrent to occupy the bandwidth.

Due to the limited upstream bandwidth video clients must have bandwidth adaptation techniques to determine the difference between congestion loss and random loss.


Figure1.Experimental Test bed