Enhancement of Data Center Transmission Control Protocol Performance in Network Cloud Environments

• Published in: Ingénierie des systèmes d'Information 2024-06, Vol.29 (3), p.1115-1123
• Main Authors: Kadhim, Qusay Kanaan, Altameemi, Atyaf Ismaeel, Abdulkader, Rasha Mahdi, Ahmed, Shaymaa Taha
• Format: Article
• Language: eng ; fre
• Subjects: Batch processing; Cloud computing; Communication; Completion time; Computer centers; Data centers; Data transmission; Internet; Performance enhancement; Performance prediction; Queuing theory; TCP (protocol); Traffic congestion
• ISSN: 1633-1311; 2116-7125
• DOI: 10.18280/isi.290329

The increasing use of cloud services in several areas has led to the growth of data-intensive applications. It is necessary to find ways to enhance the efficiency of communications within the networks of data centers to improve the performance of cloud environments. Explicit Congestion Notification (ECN) is used by Data Center TCP (DCTCP) to enhance congestion control in data center networks. The DCTCP uses ECN to assess the amount of congestion, whereas normal TCP congestion management can simply detect its presence. This paper examines DCTCP using the Random Early Detection (RED) queue management strategy. The evaluation reveals that employing Random Early Detection incurs certain costs. The RED is criticized on the one hand for both short- and medium-term connections due to longer completion time delays compared to typical DCTCP techniques. Because of ECN, DCTCP may maintain small queue sizes. However, because RED uses the average queue size, it penalizes short-lived traffic because it does not reach the bottleneck quickly. An intelligent queue management mechanism with ECN is believed to enhance DCTCP's performance in a cloud-computing environment by predicting sending rates and providing fast feedback on queue length.