[Paper Review] Communication, Computing and Caching for Mobile VR Delivery: Modeling and Trade-off

Yaping Sun, Zhiyong Chen+2 more|arXiv (Cornell University)|Apr 27, 2018

Caching and Content Delivery10 references3 citations

TL;DR

This paper proposes a joint caching and computation offloading framework for mobile VR delivery to minimize average transmission rate under latency and energy constraints. By leveraging device-side caching and computing, it analytically derives optimal trade-offs between communication, computing, and caching, showing that communication overhead can be reduced via intelligent offloading and caching, with a low-complexity greedy algorithm achieving near-optimal performance in heterogeneous scenarios.

ABSTRACT

Mobile virtual reality (VR) delivery is gaining increasing attention from both industry and academia due to its ability to provide an immersive experience. However, achieving mobile VR delivery requires ultra-high transmission rate, deemed as a first killer application for 5G wireless networks. In this paper, in order to alleviate the traffic burden over wireless networks, we develop an implementation framework for mobile VR delivery by utilizing caching and computing capabilities of mobile VR device. We then jointly optimize the caching and computation offloading policy for minimizing the required average transmission rate under the latency and local average energy consumption constraints. In a symmetric scenario, we obtain the optimal joint policy and the closed-form expression of the minimum average transmission rate. Accordingly, we analyze the tradeoff among communication, computing and caching, and then reveal analytically the fact that the communication overhead can be traded by the computing and caching capabilities of mobile VR device, and also what conditions must be met for it to happen. Finally, we discuss the optimization problem in a heterogeneous scenario, and propose an efficient suboptimal algorithm with low computation complexity, which is shown to achieve good performance in the numerical results.

Motivation & Objective

• Address the ultra-high data rate and low-latency requirements of mobile VR delivery, which strain 5G wireless networks.
• Reduce backhaul and wireless link traffic by exploiting caching and computing capabilities at mobile VR devices.
• Formulate and solve an optimization problem to minimize average transmission rate under latency and local energy consumption constraints.
• Analyze the fundamental trade-off among communication, computing, and caching in both symmetric and heterogeneous network scenarios.
• Propose an efficient suboptimal algorithm with low computational complexity for practical deployment in heterogeneous environments.

Proposed method

• Develops a framework where the MEC server handles stitching and equirectangular projection, while the mobile VR device performs rendering and optionally projection.
• Models the system with a symmetric scenario to derive a closed-form expression for the minimum average transmission rate.
• Applies dynamic programming to solve the optimal joint caching and offloading policy in the symmetric case, with complexity O(N²).
• Proposes a greedy algorithm (GA) to reduce complexity to O(N log N) with O(1) memory, using priority sorting based on rate and energy gain per unit resource.
• Introduces two-stage allocation: first, greedy caching based on the ratio $ \frac{P_i R_i^0}{I_i} $, then additional offloading based on $ \frac{R_i^0 - R_i^1}{k I_i w_i f_1^2} $.
• Extends the framework to heterogeneous scenarios where file popularity and device capabilities vary, using normalized metrics $ \bar{C}' $ and $ \bar{E}' $ for analysis.

Experimental results

Research questions

• RQ1 How can caching and computation offloading at the mobile VR device be jointly optimized to minimize average transmission rate?
• RQ2 What is the fundamental trade-off between communication, computing, and caching in mobile VR delivery under latency and energy constraints?
• RQ3 Under what conditions can communication overhead be effectively reduced by leveraging device-side computing and caching capabilities?
• RQ4 How does system heterogeneity (e.g., varying file popularity and device capabilities) affect the optimal trade-off between communication, computing, and caching?
• RQ5 Can a low-complexity algorithm achieve near-optimal performance in heterogeneous mobile VR scenarios?

Key findings

• In the symmetric scenario, the optimal joint caching and offloading policy achieves a closed-form expression for the minimum average transmission rate.
• The minimum transmission rate $ R^* $ decreases linearly with caching size $ C $ when $ \frac{N\bar{E}}{kf_1^2 I w} > C $, and remains constant otherwise.
• The transmission rate $ R^* $ first decreases and then increases with processor frequency $ f_1 $ when $ \frac{N\bar{E}}{kf_1^2 I w} > C $ and $ R^0 \geq R^1 $, indicating a non-monotonic trade-off.
• When $ R^0 < R^1 $, $ R^* $ remains unchanged with $ f_1 $, showing that computation offloading gains are saturated.
• The proposed greedy algorithm achieves good performance with $ \mathcal{O}(N\log N) $ time complexity and $ \mathcal{O}(1) $ memory, making it practical for large-scale systems.
• Numerical results confirm that the 3C trade-off in heterogeneous scenarios mirrors symmetric behavior, with diminishing returns in caching gain as file popularity increases ($ \gamma = 0.8 $).