Improving throughput of TCP flows by utilizing TCP-aware backpressure scheduling

• Main Authors: Kassim P.D.M., Muhammed A., Kadir M.F.A., Mohamed M.A., Rahiman A.R.A., Jamil N.
• Other Authors: 58706243600
• Format: Conference Paper
• Published: American Institute of Physics Inc. 2024
• Subjects: Backpressure Scheduling Algorithm; Scheduling; TCP-Aware; Transmission Control Protocol (TCP); Wireless Networks

Nowadays, wireless network technology is becoming an important thing to one's life as it provides assistance for different type of communications and thus changing the paradigm of existing data networking. Due to an increase in demands for various new applications and features that provide support for highly mobile behaviour, large volume of data transmission and a high number of users access, more and more advanced and complex technological solutions are needed in the future. One way to accommodate this, a good routing and scheduling algorithm that can improve network performance by dynamically allocating wireless resources to maximise the network throughput, accessibility fairness and latency. Backpressure algorithm is one of the techniques that can be used to achieve the maximum network throughput policies in wireless environments by routing the traffic over a multi-hop network dynamically using congestion gradient that directing the traffic around a queuing network. To avoid incompatibility issue with TCP connection, a TCP-Aware backpressure routing and scheduling algorithm was introduced to achieve the high throughput between the different numbers of flows in the multi-hop channel. In this paper, we investigate the ability of a TCP flows with different numbers of queues to perform packet transmission. Moreover, we proposed a modified TCP-Aware backpressure routing and scheduling algorithm further improve the network throughput of the existing TCP-Aware backpressure routing and scheduling algorithm. The results show the proposed algorithm outperforms the existing TCP-Aware backpressure routing and scheduling algorithm where the simulation have shown an increase of 6.04% in the network throughput. � 2023 Author(s).