Abstract
More and more scholars focus on mobile edge computing (MEC) technology, because the strong storage and computing capabilities of MEC servers can reduce the long transmission delay, bandwidth waste, energy consumption, and privacy leaks in the data transmission process. In this paper, we study the cache placement problem to determine how to cache videos and which videos to be cached in a mobile edge computing system. First, we derive the video request probability by taking into account video popularity, user preference and the characteristic of video representations. Second, based on the acquired request probability, we formulate a cache placement problem with the objective to maximize the cache hit ratio subject to the storage capacity constraints. Finally, in order to solve the formulated problem, we transform it into a grouping knapsack problem and develop a dynamic programming algorithm to obtain the optimal caching strategy. Simulation results show that the proposed algorithm can greatly improve the cache hit ratio.
1. Introduction
As 5G technology rolls out globally, the demand for ultra-high-definition (UHD) video and Virtual Reality (VR) content is skyrocketing. Delivering this bandwidth-intensive content from centralized cloud servers often results in high startup latency, significant backhaul network congestion, and excessive energy consumption. Mobile Edge Computing (MEC) addresses these issues by placing storage and computing resources at the network edge, such as base stations, closer to end-users.
However, MEC servers are constrained by limited storage capacity. The primary challenge lies in determining which videos to store and in what format (e.g., High-Definition vs. Standard-Definition) to maximize user satisfaction. Traditional caching algorithms like Least Recently Used (LRU) or Least Frequently Used (LFU) often fail to account for user-specific preferences and the unique multi-representation nature of modern video streams.
2. System Model and Methodology
The proposed system utilizes a hierarchical architecture where MEC servers are positioned between the central cloud and the user terminals. By leveraging the computational power of MEC servers for real-time transcoding, the system can serve multiple quality requests from a single high-quality cached file, effectively acting as a storage multiplier.
The authors propose a three-step optimization approach:
- Comprehensive Request Probability Calculation: Request probability is modeled by integrating global video popularity (Zipf distribution), individual user preference factors (adjusting for content categories like news or sports), and representation characteristics (quality-specific demand).
- Joint Optimization with Transcoding: The model assumes the MEC server can transcode HD video into SD on-the-fly. This "one-to-many" capability allows the server to satisfy SD requests without dedicated storage, optimizing the use of limited edge capacity.
- Grouping Knapsack Problem (GKP) Formulation: The cache placement problem is mapped to a GKP, where each video represents a group and its various quality levels are items within that group. A Dynamic Programming (DP) algorithm is then employed to find the optimal set of videos that maximizes the total hit ratio within storage constraints.
3. Experimental Setup
The performance of the proposed scheme was evaluated in a simulated environment with the following parameters:
- Content Library: 21 videos across five categories (news, scene, sport, traffic, person).
- User Base: 900 users distributed across 4 base stations.
- Traffic Model: Popularity followed a Zipf distribution (decay parameter $\gamma = 0.6$), and requests followed a Poisson distribution (intensity $\lambda = 0.9$).
- Storage: Video sizes ranged from 3 to 65 storage units.
4. Results and Discussion
The proposed algorithm was benchmarked against LFU, LRU, and Weighted Greedy Dual Size Frequency (WGDSF) methods. The results demonstrate a significant performance advantage for the hit-ratio-driven approach.
Key findings include:
- Superior Hit Ratio: The proposed scheme consistently outperformed competitors, achieving a hit ratio approximately 10% higher than the next best algorithm (WGDSF).
- Storage Efficiency: As shown in the performance curves, the proposed method reaches a 100% hit ratio at lower storage capacities compared to traditional methods, indicating better utilization of edge resources.
- QoS Improvements: Higher hit ratios directly correlate with reduced backhaul load and lower startup delays, enhancing the overall Quality of Service (QoS) for VoD users.
- Content Sensitivity: The algorithm successfully prioritized content based on localized user preferences, such as favoring "Sports" videos in clusters with a high statistical preference for that category.
5. Conclusion and Future Work
This research demonstrates that integrating user preferences and transcoding capabilities into edge caching strategies significantly improves network efficiency. While the current model assumes sufficient computational resources for transcoding, future work should address the trade-offs between transcoding latency and storage savings. Additionally, moving from static daily popularity updates to real-time, dynamic popularity tracking will be essential for handling viral content in live network environments.
Summary updated