Performance and energy consumption optimization of ternary optical computers based on the M/G/1 queuing model.

Area of Science:

• Computer Science
• Electrical Engineering
• Operations Research

Background:

• Ternary Optical Computers (TOC) face challenges in balancing quality-of-service (QoS) and energy consumption.
• Existing service models do not adequately address these dual objectives in TOC environments.

Purpose of the Study:

• To propose and analyze a novel three-phase service model for TOC systems.
• To optimize energy consumption and QoS through integrated queuing and vacation mechanisms.
• To provide theoretical insights and practical guidelines for TOC dual-objective optimization.

Main Methods:

• Integration of an M/G/1 queuing system with an N-policy and multiple vacations into the TOC service framework.
• Stochastic-decomposition methods for precise steady-state queue length analysis of the third phase.
• Approximate analysis for the first and second phases to balance computational complexity and accuracy.
• Application of Little's Law to derive the overall system's sojourn time.
• Utilization of the Renewal Reward Theorem to model TOC energy consumption and identify optimal parameters.

Main Results:

• The proposed model significantly reduces energy consumption in TOC systems.
• Satisfactory system performance (QoS) is maintained alongside energy savings.
• Numerical experiments successfully identified an optimal threshold (N) for the N-policy.

Conclusions:

• The novel three-phase service model effectively addresses the dual objectives of QoS and energy consumption in TOC.
• The study offers valuable theoretical contributions and practical recommendations for optimizing TOC performance.
• The integrated approach provides a robust framework for future research in energy-efficient computing systems.