Dynamic Cost-Aware Routing of Web Requests

Area of Science:

• Computer Science
• Network Engineering
• Energy Management

Background:

• Next-generation networks face increasing demand for ubiquitous services and massive traffic from applications like the Internet of Things.
• Redundant, geographically distributed servers are essential for reliable web services, but lead to significant energy costs for operators.
• Optimizing server deployment and request routing is crucial for managing operational expenses.

Purpose of the Study:

• To propose a novel, cost-aware method for routing web requests across distributed servers.
• To leverage spatial and temporal variations in electricity prices and network topology for cost optimization.
• To reduce energy consumption and operational costs for web service providers.

Main Methods:

• Development of a cost-aware routing method utilizing a learning automaton.
• Implementation of per-request decision-making for efficient computation.
• Simulation and testbed measurements to evaluate the proposed method against standard algorithms.

Main Results:

• Demonstrated significant cost reductions in web request routing.
• Achieved these cost savings with minimal impact on overall performance.
• Validated the effectiveness of the learning automaton approach for dynamic routing decisions.

Conclusions:

• The proposed cost-aware routing method effectively reduces energy costs for web services.
• Learning automaton-based routing offers a computationally efficient alternative to global optimization.
• This approach is vital for sustainable and cost-effective operation of large-scale web infrastructures.