Refresh Rate-Based Caching and Prefetching Strategies for Internet of Things Middleware

Area of Science:

• Computer Science
• Internet of Things (IoT)
• Distributed Systems

Background:

• Context Management Platforms (CMPs) are crucial IoT middleware enabling connectivity across diverse IoT systems.
• Efficient processing and monetization of IoT data streams are essential for large-scale IoT ecosystems.
• Caching transient context information incurs significant computational and network costs.

Purpose of the Study:

• To develop a method for modeling and predicting context metrics and operational costs in IoT data exchange.
• To propose strategies for minimizing context management costs through efficient caching and prefetching.
• To maximize the profit of CMP operations under multiple Service Level Agreements (SLAs).

Main Methods:

• Utilizing Service Level Agreements (SLAs) to quantify service imperfection and link it to costs.
• Developing mathematical models for context caching and prefetching strategies.
• Proposing a novel method to determine the optimal rate of IoT data prefetching and caching.

Main Results:

• Demonstrated that efficient caching and prefetching strategies can significantly minimize context management costs.
• Showcased how proactive caching strategies maximize CMP profits when multiple SLAs are defined.
• Achieved a Root Mean Square Percentage Error of up to 0.0016 in profit estimation against simulation results.

Conclusions:

• The proposed method effectively optimizes IoT data caching and prefetching to reduce operational costs.
• The mathematical models and strategies presented enable profit maximization for CMPs.
• The model's validity is confirmed through experimental scenarios with a t-test value tending to 0.