Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Community
Ask the community for help and clear up your study doubts
Discover the best universities in your country according to Docsity users
Free resources
Download our free guides on studying techniques, anxiety management strategies, and thesis advice from Docsity tutors
Material Type: Project; Class: 4005 - Adv Computer Networks; Subject: Graduate Computer Science; University: Rochester Institute of Technology; Term: Unknown 1989;
Typology: Study Guides, Projects, Research
1 / 25
This page cannot be seen from the preview
Don't miss anything!
We performed our investigation by choosing three papers which were presented at various conferences. In this section, we would be explaining the analysis of the three research papers
The IP protocol suite does not guarantee Quality of Service (QoS) and prioritized transmission. However, in case of transmission of video and voice, there are constraints like delay and packetizing which need to be taken into consideration. There are several critical issues which exist for real time communication for the IP network such as reduction in bandwidth, packet delay and echo. There are various challenges that arise in real time communication over IP network. Some of them are:
The authors of the paper have proposed an idea and guarantees performance in real time video communication with error resilience. The assumptions made in the paper are as follows :
The novel contributions made by the paper are discussed in this section:
a) Channel Establishment : The channel establishment phase is used to determine if the client’s request can be accepted by the network in its current load conditions. It is achieved using Open Shortest Path First (OSPF) .The route is scrutinized to cand their heck enough resources. A channel establishment message follows a particular path and tests are performed. The results are forwarded to the destination which includes the minimum delay bound a node can offer and the buffer space the node can offer. If one of tests fails at a node, a negative reply message is sent back to the source and all the resources that are held by the node are released.
b) Backpressure Rerouting The backpressure rerouting takes place when the traffic pattern changes and during congestion. Each node (say node x) maintains information about all the intermediate nodes (source to node x) along with the successor node and their delay requirements. On receiving a backpressure packet, node x tries to find an alternative path by issuing a similar packet which was issued by the source for channel establishment. This packet contains information about the source to node x along with the delay characteristics. A neighbor table is maintained which records the delay and the expiry time for a node to reach its neighbors.
My Neighbors
Estimated Delay (ms)
Expiry Time (ms)
Node 23 17 50
Node 9 234 50
Fig 1 : Neighbor Table
The Encoder and Decoder are used for detecting errors. The Feedback loop is utilized to recover from errors. There are various video compression algorithms which are used:
We have simulated a sparsely connected network. We have taken the idea of backpressure rerouting in order to perform the rerouting mechanism when congestion happens in the network. We have also implemented the rerouting mechanism with Forward Error Correction (FEC) in order to check the error resilience of the network.
Routers respond to congestion by dropping packets. However, dropping the packets indiscriminately would encourage the senders to increase their sending rate due to which it becomes difficult to control congestion. Fair Queuing is one of the congestion control mechanisms which have been proposed. In this mechanism, separate queues are maintained for each source. Whenever congestion happens, the packets from the high volume sources are dropped.
The paper illustrates the drawbacks of the various current techniques which have been designed for performing congestion control. Some of the techniques described are:
The above mentioned mechanisms have its own drawbacks. The techniques described drop packets at random, due to which the sender is encouraged to send more packets and thus clogging of the network takes place.
The paper proposes a packet dropping mechanism based on game theory. Some of the parameters used in the proposed protocol are:
Every time a packet arrives at the queue, fair congestion mechanism takes place, where the packet is scrutinized at the tail of the queue (arrival) and departure of the queue (depart).
The steps to be performed are as follows:
Packet Arrival
Packet Departure
The authors then go on to discuss different packet loss patterns and the relationship between static FEC, Throughput and Utility for each of them. This helps the authors towards building their analysis which helps them to conceive the adaptive FEC approach. The authors then extrapolate a relationship between packet loss p and FEC overhead ψ. Based on this relationship they propose a simple proportional controller that adjusts ψ based on measured packet loss.
where i represents the flow number, pi(n) is the average packet loss in the ith flow during
i. The measurement of pi is coupled with congestion control and should be provided via
2 is satisfied.
Now they say that suppose the probability pi(n) converges to some stationary point p*, then the overhead rate in the stationary state becomes
Now the authors state that the task of achieving high utility becomes the matter of
Thus armed with the knowledge of equations (1) and (2) we can effectively achieve adaptive FEC to maintain high utility for constantly changing packet loss ratios. And this is what we use in the third facet of our project. The last part of this paper shows us the simulation results for the experiments the authors carried out. Basically they go on to prove that proper control of FEC overhead can significantly improve the utility of received video over lossy channels and have better performance than other best-effort networks with FEC-protection.
In this section, the class diagrams for the simulation software are illustrated. We have designed two class diagrams: one for the Fair and Efficient dropping technique and one for the Backpressure rerouting technique. The classes in the simulation software are
For Linux/MAC OS User: (Using Bash shell)
*If Required: Set the PATH and CLASSPATH to include the following from the bash command
line or add the following lines to your .bashrc file in the user's home folder.
$ export PATH=$PATH:/home/user/
$ export CLASSPATH=.:/home/user/
It is expected that the user change the necessary paths to the folders and the java (jdk) version as
well
to suit his/her computer configurations. It is recommended that the Java Development Kit version be
jdk1.6.0_
Download the 2 JAR files <ccrushers.jar> and <congestionc.jar>
ccrushers.jar is the Back Pressure Rerouting Program with FEC inbuilt into it.
congestionc.jar is the Fair and Efficient Dropping Algorithm with FEC inbuilt
Move the file into the folder
command from the bash shell.
$ jar -xvf ccrushers.jar
Then move into another folder
$ jar -xvf congestionc.jar
This will copy all the source files and configuration files into the respective folder. Now compile
the java source files in each of the folders using the following command on the bash shell.
$ javac *.java
Now all the java .class files have been created and the program is set to run.
Once the class files are made available to the user the user is expected to run the programs thus:
$ java SimulationLauncher
seed : is the seed for the random number generators. NetworkInputFile : is all the configurations that the program would need its format is as follows:
NOTE: DO NOT USE THIS FILE AS INPUT FOR SIMULATION
0:0 3:0 (Source Host) - Can be more
We have collected data from the simulation by varying the arrival rate in every simulation. The
following data has been collected.
Fig 3. Data collected of Drop mechanism with FEC and without FEC
