Analysis of Congestion Control Techniques for Time Critical Applications

Abstract

Social media video applications, such as TikTok, require smooth and uninterrupted data transmission. These
applications are time-sensitive and could not tolerate long delays in transmission caused by data transmission
protocols. For example, the congestion control and reliability check protocols, TCP and UDP, are used in today’s
Internet. TCP is a reliable transport layer protocol with congestion control mechanism which delivers the data in
an ordered manner and retransmits the data in case of errors. TCP needs improvement when used for applications
in which reliability could be compromised for high performance. UDP is suitable for time-sensitive applications, but
it has no mechanism to keep a smooth transmission in case of congestion. Datagram congestion control protocol
(DCCP) has been developed to overcome the weaknesses of TCP and UDP with more control on the congestion
and the timely delivery of data. It delivers the data in time and also has congestion control mechanism. This paper
compares the performance of advanced congestion control techniques of DCCP, such as CCID 2 and CCID 3,
over different networks through simulations. The proposed simulation networks are configured with a high speed
bandwidth and random link failures. The results show that CCID 2 (TCP-like) is better in dealing with network
congestion in a 5-node scenario. Whereas, on a 20-node scenario and a link failure scenario, CCID 3 (TFRC)
outperforms CCID 2 and TCP.