Accurate Server Resource Allocation Using Network Performance Prediction

Read full text Summarise with VIRAT
Authors: Lingala Mithun Kumar Reddy, Pusalapati Chandu, Pulipakula Akhil Kumar
Cited by: 0
References: 12

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Server resource allocation is an important problem for cloud computing and distributed systems. Modern data center services experience highly variable loads, which are combined with networks with complex dynamics. Without optimal resource allocation, improper resource allocation leads to poor resource efficiency, longer transaction times, decreased quality of service (QoS), and higher costs. This has proved to be insufficient for contemporary, timeliness-sensitive, and large-scale data centers. This research proposed a smart resource allocation technique called smart SRCA (Server Resource Capacity Allocation) with the aid of network performance prediction. This method effectively utilized historic and real time network data to provide appropriate resource allocation. In this study, we utilize server parameters and network performance metrics such as latency, average bandwidth utilization, packet loss, throughput, CPU load and current number of active users. These metrics help in training the proposed models for network performance prediction (NPP - Network Performance Prediction) and workloads prediction. Based on these models, appropriate server capacity can be dynamically allocated for CPU, memory, storage and network bandwidth, which are all factors affecting service quality provided to users. The results of the performance evaluation confirmed that incorporating the predictive analytics techniques into resource allocation not only improved the data center's capability in making smart decisions about the resources, also assisted in increasing data center performance metrics such as average throughput, resource utilization, transaction speed and reliability. In future, the users can use the deep learning models for prediction with a real time SRCA implementation on edge computing architecture and implement using Software-Defined Networking (SDN) protocols.

Fields of Study & Subjects

Cloud Computing and Resource Management Software-Defined Networks and 5G Energy Efficiency in Computing

Publication details

Journal / source
Journal of Information Technology and Digital World
Volume
8
Issue
1
Pages
15–25
Publication date
20 March 2026
Language
EN