Novel Record Replacement Algorithm and Architecture for QoS Management over Local Area Networks.

Area of Science:

• Computer Engineering
• Network Management
• Artificial Intelligence

Background:

• Effective Quality-of-Service (QoS) management is crucial for modern virtual local area networks (VLANs).
• Existing methods for real-time QoS prediction can be computationally intensive and may not adapt well to dynamic network conditions.
• The need for intelligent, self-aware systems that can optimize network performance with low overhead is increasing.

Purpose of the Study:

• To present an effective System-on-Chip (SoC) designed for smart QoS management within VLAN environments.
• To accelerate the prediction of service delivery quality using hardware acceleration.
• To enable adaptive and efficient bandwidth usage tracking for services.

Main Methods:

• Implementation of the SoC using a field programmable gate array (FPGA) for hardware acceleration.
• Utilization of the general regression neural network (GRNN) algorithm for quality prediction.
• Development of a novel record replacement algorithm to dynamically update service profiles and track bandwidth usage.

Main Results:

• The developed SoC effectively accelerates the delivery quality prediction process.
• The GRNN, combined with the novel record replacement algorithm, accurately tracks time-varying service bandwidth usage.
• Experimental validation demonstrates the SoC's capability for self-aware QoS management.

Conclusions:

• The proposed SoC offers a powerful solution for intelligent QoS management in virtual LANs.
• The system achieves self-aware QoS management with significantly low computation costs.
• This approach enhances the performance and reliability of applications operating over virtual LANs.