Predictive Pre-fetching Service for Mobile Content Delivery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Mobile users experience significant latency and low Quality of Experience (QoE) when accessing content, such as video, over cellular networks due to network congestion and inefficient content delivery, as existing content delivery networks (CDNs) are not optimized for mobile networks and often result in packet drops and retransmissions during peak traffic hours.
Innovation Solution
Implementing a predictive pre-fetching service within CDNs that allows mobile devices to subscribe to content and pre-fetch it based on user preferences, using a distributed predictive pre-fetching function across CDN server clusters, with a centralized back-end infrastructure and front-end servers positioned near mobile core networks, enabling content to be downloaded in the background during low-load periods and cached for on-demand viewing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If CDN delivers content interactively based on user requests, then content delivery is efficient in terms of network resource usage, but user perceived latency is high and Quality of Experience is low
Solution Approach 1:
The system performs preliminary actions by predicting user content requests using machine learning models and pre-fetching content to edge servers before users actually request it. This eliminates user perceived latency while optimizing network resource usage by pre-loading content during off-peak periods and only transferring what is predicted to be needed.
2Speed
If CDN edge server is positioned closest to mobile network IP gateway, then network delivery path is optimized, but content delivery quality remains unsatisfactory due to network congestion and packet drops
Solution Approach 1:
The system segments content delivery into multiple stages: prediction at cloud centers, pre-fetching to regional edge servers, and final delivery to users. This segmentation allows content to be delivered in smaller, more reliable chunks while maintaining high speed through optimized pathways and reducing the impact of network congestion and packet drops.
3Loss of energy
If adaptive bit rate streaming is used to compensate for network congestion, then network resource usage is reduced, but Quality of Experience deteriorates due to lower video quality
Solution Approach 1:
The system pre-fetches high-quality content to edge servers before user requests, allowing delivery of higher bit-rate video without network congestion issues. This preliminary action enables quality video delivery while optimizing network resource usage through advance preparation during off-peak periods.
Data Source
AI summary
A content delivery network (CDN) is enhanced to enable mobile network operators (MNOs) to provide their mobile device users with a content prediction and pre-fetching service. Preferably, the CDN enables the service by providing infrastructure support comprising a client application, and a distributed predictive pre-fetching function. The client application executes in the user's mobile device and enables the device user to subscribe to content (e.g., video) from different websites, and to input viewing preferences for such content (e.g.: “Sports: MLB: Boston Red Sox”). This user subscription and preference information is sent to the predictive pre-fetching support function that is preferably implemented within or across CDN server clusters. A preferred implementation uses a centralized back-end infrastructure, together with front-end servers positioned in association with the edge server regions located nearby the mobile core network. The predictive pre-fetch service operates on the user's behalf in accordance with the user preference information.