Congestion Control Mechanism for TCP in Data Centered Networks Using Multithreading

Transport Control Protocol (TCP) incast congestion happens in high-bandwidth and low-latency networks once multiple synchronized servers send data to identical receiver in parallel. TCP congestion could severely degrade the performance in many-to-one traffic pattern applications like MapReduce and search. The main limitation of TCP is that congestion in network which resulted into huge data loss, more waiting time etc. However existing methods focuses on either on the reduction of the waiting time for packet loss recovery using the faster retransmissions or controlling switch buffer occupation to prevent overflow with ECN as well as modified TCP on both the sender and receiver sides. This system is mainly focusing on presenting efficient solution which can be able to prevent packet loss before the incast congestion condition in network rather than packet loss recovery after actual packet loss in existing cases. This proposed method presented in this system is nothing but congestion avoidance in TCP based networks. The system presents Cloud Incast Congestion Control for TCP (CICT) scheme on the receiver side using the concepts of multi threading to manage parallelism concepts of data communication in data center networks. This method not only reduces the costs required for router or switch modifications but also minimizes the working on sender and receiver nodes.