Truthful online double auction based dynamic resource provisioning for multi-objective trade-offs in IaaS clouds

Area of Science:

• Computer Science
• Cloud Computing
• Distributed Systems

Background:

• Existing Infrastructure as a Service (IaaS) cloud auction designs are limited to simple, single-objective, offline scenarios, often neglecting user privacy and one-sided interactions.
• Current mechanisms struggle with dynamic resource provisioning, time-varying user requests, and server maintenance schedules in cloud data centers.

Purpose of the Study:

• To propose a realistic online auction mechanism for IaaS clouds that addresses elasticity and server maintenance.
• To develop a truthful double auction technique that balances multi-objective trade-offs including energy consumption, revenue generation, and performance.
• To ensure fairness, efficiency, and computational feasibility in cloud resource allocation and pricing.

Main Methods:

• Developed an online truthful double auction framework for IaaS cloud environments.
• Implemented a weighted bipartite matching algorithm for determining winning bids and allocating resources.
• Utilized a Vickrey-Clarke-Groves (VCG) driven algorithm for calculating payments for awarded bids.

Main Results:

• The proposed mechanism effectively balances multi-objective trade-offs between energy, revenue, and performance in IaaS clouds.
• Demonstrated truthfulness, heterogeneity, economic efficiency, and individual rationality through theoretical analysis and simulations.
• Achieved polynomial-time computational complexity for resource allocation and payment calculations.

Conclusions:

• The novel online truthful double auction significantly improves performance in IaaS clouds compared to existing methods.
• The mechanism provides a more realistic and robust approach to resource provisioning in dynamic cloud environments.
• This research offers a promising solution for efficient and fair resource management in cloud data centers.