CT-PIFO: A Congruity Tree Based Packet Scheduler to Improve Flow Completion Time

Abstract

5G technology is driving significant innovations and shaping a future of pervasive connectivity by enabling the seamless connection of a large number of devices within a network. As the number of interconnected devices grows and data demand intensifies, the need for efficient packet scheduling becomes essential in ensuring consistent Quality of Service (QoS). Efficient scheduling mechanisms must be capable of dynamically adjusting to fluctuating network conditions, ensuring optimal resource allocation, minimizing latency, and sustaining high throughput. To this end, we propose CT-PIFO-a congruity metric-based packet scheduling mechanism to provide PIFO-like performance. Packets are stored in the priority queues based on the congruity metric, which guides a ranked packet to the appropriate queue. Our proposed design substantially improves the performance over the state-of-the-art models such as SPPIFO. Simulation results show that our design provides ~10% better flow completion time as compared to SP-PIFO and 2-3× reduction in packet drop. Our system better approximates the PIFO by falling within 2-3% of PIFO's flow completion time. © 2025 Elsevier B.V., All rights reserved.