Performance analysis of a novel transport protocol without congestion control in Cisco-based testbeds

OData support
Supervisor:
Dr. Sonkoly Balázs
Department of Telecommunications and Media Informatics

Congestion control was one of the most important control mechanisms managed

by the TCP from the beginning of the internet. However, some new versions have been developed because of the changing of the internet. These versions have the traditional TCP congestion control changed. They could increase the utilization of the resources by these changing. The new mechanisms could be used in some today’s environment. I said the fact that only some environment because no other version could give universial solution for the problem. It seems the congestion control by TCP will never be suitable to serve any claim. The researchers were inspired by the shortcomings of the TCP to work out totally new methods. Their new idea was not to use any congestion control in the network.

In the first part of my thesis I present the traditional TCP protocol and thus I mention the fault and the mistakes, too. After that, I review the versions proposed to the high-speed network environment and the solution of these at TCP. In the next part I present the protocol without congestion control as the new idea, which is called P7 protocol, too. The endpoints, using this protocol, send data at maximum speed to the network. Therefore, the congestion and the packet loss are inescapable what can be fixed by effective error-correcting codes. I show this coding process and furthermore the working and the header of P7. Turning to Cisco devices I present the possibilities of QoS (Quality of Sevice) settings and their theoretical background. After that, I create topologies using routers and switches in order to test the capability of their data transfer in the case of TCP and P7 protocol. I plot the topologies on graphs and I describe their create progress and settings in details. There is an example for meaning of the settings, you can get totally different resaults using variant QoS configures. The second part of my thesis I review the programs used for my measuring and the resaults of them. I survey the serviceable parameters at the programs and I explain the cause of my chioce. I clearly plot the resaults and I explain them in every case. Finally I compare the two tested protocol based on the graphs. At the end of my thesis I do a summary and I write the possible future of P7.

Downloads

Please sign in to download the files of this thesis.