DeepAuC: Joint deep learning and auction for congestion-aware caching in Named Data Networking.

Area of Science:

• Computer Science
• Network Engineering

Background:

• Internet data traffic has grown exponentially, increasing content consumption.
• Content retrieval from providers causes network congestion, delays, and high latency.
• Existing solutions struggle to manage escalating network demands effectively.

Purpose of the Study:

• To propose a novel approach for congestion-aware caching in Named Data Networking (NDN).
• To minimize content downloading delays and prevent network congestion.
• To optimize content caching strategies for Internet Service Providers (ISPs) and Content Providers (CPs).

Main Methods:

• A deep learning model is developed to predict future traffic on transit links using historical ISP network data.
• A predictive caching model is implemented to cache high-demand content, preventing future congestion.
• An auction mechanism is designed to acquire paid content at an optimal price.

Main Results:

• The proposed approach effectively prevents network congestion on transit links.
• Content downloading delays are significantly minimized.
• The strategy enhances profitability for both ISPs and CPs.

Conclusions:

• The joint deep learning and auction-based approach offers an effective solution for congestion-aware caching in NDN.
• This method improves network performance by reducing latency and preventing congestion.
• It presents a mutually beneficial economic model for network infrastructure providers and content creators.