[Paper Review] Age-of-Information for Computation-Intensive Messages in Mobile Edge Computing

Qiaobin Kuang, Jie Gong+2 more|arXiv (Cornell University)|Jan 7, 2019

Age of Information Optimization11 references7 citations

TL;DR

This paper studies age-of-information (AoI) for computation-intensive messages in mobile edge computing (MEC), proposing a two-node tandem queuing model to compare local computing (user-side) and remote computing (MEC server-side) under zero-wait policy. It derives closed-form average AoI expressions and shows that remote computing significantly reduces AoI when MEC computing capacity is high or transmission rate is optimal, with an optimal data rate minimizing AoI for remote execution.

ABSTRACT

Age-of-information (AoI) is a novel metric that measures the freshness of information in status update scenarios. It is essential for real-time applications to transmit status update packets to the destination node as timely as possible. However, for some applications, status information embedded in the packets is not revealed until complicated data processing, which is computational expensive and time consuming. As mobile edge server has sufficient computational resource and is placed close to users, mobile edge computing (MEC) is expected to reduce age for computation-intensive messages. In this paper, we study the AoI for computation-intensive data in MEC, and consider two schemes: local computing by user itself and remote computing at MEC server. The two computing models are unified into a two-node tandem queuing model. Zero-wait policy is adopted, i.e., a new message is generated once the previous one leaves the first node. We consider exponentially distributed service time and infinite queue size, and hence, the second node can be seen as a First-Come-First-Served (FCFS) M/M/1 system. Closed-form average AoI is derived for the two computing schemes. The region where remote computing outperforms local computing is characterized. Simulation results show that the remote computing is greatly superior to the local computing when the remote computing rate is large enough, and that there exists an optimal transmission rate so that remote computing is better than local computing for a largest range.

Motivation & Objective

• To analyze the age-of-information (AoI) for computation-intensive messages in mobile edge computing (MEC), where data processing is time-consuming and resource-intensive.
• To compare two computing schemes—local computing at the user and remote computing at the MEC server—under a unified two-node tandem queuing model.
• To characterize the region where remote computing outperforms local computing in terms of average AoI.
• To investigate the impact of key system parameters such as packet size, required CPU cycles, transmission rate, and MEC computing capacity on AoI performance.

Proposed method

• Modeling the system as a two-node tandem queuing network: user-to-MEC transmission followed by processing at the MEC server.
• Applying the zero-wait policy, where a new message is generated immediately after the previous one departs the first node.
• Assuming exponentially distributed service times and infinite queue size, enabling the second node to be modeled as an FCFS M/M/1 queue.
• Deriving closed-form expressions for average AoI under both local and remote computing schemes using stochastic process analysis.
• Using stochastic hybrid systems (SHS) to analyze the age dynamics and derive stationary age distributions.
• Characterizing the performance region where remote computing yields lower average AoI than local computing based on system parameters.

Experimental results

Research questions

• RQ1 Under what conditions does remote computing at the MEC server yield lower average AoI than local computing at the user device for computation-intensive messages?
• RQ2 How does the transmission rate affect the AoI trade-off between local and remote computing?
• RQ3 What is the optimal data rate that minimizes the average AoI in remote computing, and how does it depend on system parameters?
• RQ4 How do the required number of CPU cycles and MEC server computing capacity influence the AoI performance in remote computing?
• RQ5 Is there a threshold packet size beyond which local computing becomes preferable to remote computing?

Key findings

• Remote computing outperforms local computing when the MEC server's computing capacity is sufficiently high or the transmission rate is optimally chosen, especially for large required CPU cycles.
• An optimal transmission rate exists that minimizes the average AoI for remote computing, and this rate maximizes the performance gain over local computing.
• When the required number of CPU cycles exceeds 7000 Megacycles, the average AoI for remote computing increases sharply due to queue buildup, approaching infinity as the server utilization ρs approaches 1.
• For small packet sizes (e.g., 0.5 Mbits), remote computing outperforms local computing up to a transmission rate of approximately 1.64 Mbits, while for larger packets, the cross-over point shifts, and local computing becomes preferable at higher rates.
• The average AoI for local computing is independent of data rate and packet size, as it depends only on local processing speed.
• As MEC computing capacity increases, the average AoI for remote computing decreases and eventually converges to 2/μt, the theoretical minimum, with a significant performance gain when the required CPU cycles are large.