Scholarlink

<p>The fast development of cloud computing, IoT, artificial intelligence, and data-intensive applications has improved the necessity for efficient resource management in 5G network slicing and large-scale data center networks. But traditional resource allocation and load-balancing methods operate independently, resulting in inefficient resource utilization, congestion, increased latency, reduced throughput, and SLA violations under dynamic network conditions. The rising demand for ultra-low latency and high reliability in applications such as autonomous vehicles, smart cities, and cloud-native services motivates the development of intelligent and adaptive network management solutions. To address these challenges, this research suggests a combined framework that syndicates resource allocation and load balancing within a unified architecture. The framework employs machine learning, optimization techniques, SDN, and NFV technologies to vigorously allocate resources and distribute workloads based on real-time network conditions. The main objective is to develop resource utilization, decrease latency, increase throughput, ensure QoS compliance, rise scalability and reliability, and accomplish efficient network performance. The methodology incorporates ACALF for intelligent traffic prediction, LRO-DLB for load balancing, NSV-DO for dynamic slicing, and SDN-LALB for latency-aware routing, assessed using CIC-DDoS2019 and network traffic datasets. Results show strong performance improvements: latency reduced from 55.2 ms to 32.8 ms (ACALF) and overall to 19.6 ms, throughput increased up to 945 Mbps, congestion dropped from 34.5% to 18.2%, and SLA compliance improved to 99.1%. Energy efficiency also improved, reducing consumption from 520 kWh to 390 kWh. Future scope includes integrating federated learning, real-time edge AI, and 6G-compatible autonomous network optimization for scalable ultra-low-latency systems.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Synergistic Resource Allocation and Load Balancing Techniques for Network Slicing in 5G and Large-Scale Data Center Networks

  • 期刊论文

  • 发表日期：

    2026年7月21日

查看全文
  • S. Varadhaganapathy,
  • Tukaram N. Sawant,
  • Abilash Radhakrishnan,
  • Dani Jermisha Railis,
  • J. Kavitha

摘要

The fast development of cloud computing, IoT, artificial intelligence, and data-intensive applications has improved the necessity for efficient resource management in 5G network slicing and large-scale data center networks. But traditional resource allocation and load-balancing methods operate independently, resulting in inefficient resource utilization, congestion, increased latency, reduced throughput, and SLA violations under dynamic network conditions. The rising demand for ultra-low latency and high reliability in applications such as autonomous vehicles, smart cities, and cloud-native services motivates the development of intelligent and adaptive network management solutions. To address these challenges, this research suggests a combined framework that syndicates resource allocation and load balancing within a unified architecture. The framework employs machine learning, optimization techniques, SDN, and NFV technologies to vigorously allocate resources and distribute workloads based on real-time network conditions. The main objective is to develop resource utilization, decrease latency, increase throughput, ensure QoS compliance, rise scalability and reliability, and accomplish efficient network performance. The methodology incorporates ACALF for intelligent traffic prediction, LRO-DLB for load balancing, NSV-DO for dynamic slicing, and SDN-LALB for latency-aware routing, assessed using CIC-DDoS2019 and network traffic datasets. Results show strong performance improvements: latency reduced from 55.2 ms to 32.8 ms (ACALF) and overall to 19.6 ms, throughput increased up to 945 Mbps, congestion dropped from 34.5% to 18.2%, and SLA compliance improved to 99.1%. Energy efficiency also improved, reducing consumption from 520 kWh to 390 kWh. Future scope includes integrating federated learning, real-time edge AI, and 6G-compatible autonomous network optimization for scalable ultra-low-latency systems.