| Student | (visible for staff only) |
| Supervisor | Xiaoyuan Gu |
| Matthias Dick | |
| Professor | Prof. Dr.-Ing. Lars Wolf |
| IBR Group | CM (Prof. Wolf) |
| Type | Diploma Thesis |
| Status | finished |
Motivation: DCCP, the Datagram Congestion Control Protocol, is a new transport protocol currently being standardized by IETF that provides a congestion-controlled flow of unreliable datagrams. Recent years have seen dramatic increases of the use of multimedia applications on the Internet. Delay-sensitive applications, such as streaming media and video telephony, typically prefer timeliness to reliability. These applications have historically used UDP for transport and implemented their own congestion control mechanisms - a difficult task, or no congestion control at all. This can easily cause compatibility problems or congestion collapse in the networks. DCCP makes it easy to deploy these applications without risking congestion collapse. It aims to add to a UDP-like foundation the minimum mechanisms necessary to support congestion control. In short, DCCP is an unreliable transport protocol like UDP, but it has congestion control like TCP. The protocol can be extended by adding new congestion control algorithms, e.g. TCP-like congestion control or TCP-Friendly Rate Control (TFRC) as profiles, in order to customize the congestion control for applications with different characteristics. It is highly promising that DCCP will become the de facto standard for multimedia rich content delivery over IP-based networks. On the other hand, as DCCP provides many features and options to users, and has been criticized for the resulting complexity. Therefore a simplified version of DCCP, DCCP-Lite, has gained pretty much attention. OPNET Modeler is a leading industry-strength network modeling and simulation environment which enables the design and study of all major communication networks, devices, protocols, and applications with unmatched flexibility. With its object-oriented modeling approach and graphical editors, OPNET Modeler visualizes the structure of actual networks and network components, and the system designed intuitively maps to the model. Productivity and efficiency can be greatly boosted by leveraging the specialized editors, analysis tools, and off-the-shelf models. Due to its open source, new research proposals can be easily implemented, simulated, evaluated and tested, providing a good understanding of the simulation kernel. Tasks: In this project, the student will implement a DCCP-lite module for OPNET Modeler. This is done through first to have a good understanding of DCCP and DCCP-lite, and to study the simulation workflow in OPNET Modeler. Both congestion control mechanisms, namely TCP-like Rate Control and TCP-friendly Rate Control will be implemented. Existing implementations of DCCP will be used as references during this phase. This is followed by the simulation of the implemented protocol, the performance evaluation of DCCP-lite, the investigation of the bandwidth fairness between DCCP-lite flow and TCP flow within a common channel to verify the correctness of theories, and to seek possible optimizations for DCCP-lite. Prerequisite: Good knowledge of computer networks, and programming skills in C/C++ | |
