Are Bus-Mounted Edge Servers Feasible?

Placement of edge servers is the prerequisite of provisioning edge computing services for Internet of Vehicles (IoV). Fixed-site edge servers at Road Side Units (RSUs) or base stations are able to offer basic service coverage for end users, i.e., vehicles on road. However, the server locations and capacity are fixed after deployment, rendering their inefficiency in handling spationtemporal user dynamics. Mobile servers such as buses, on the other hand, have the potential of adding computation elasticity to such system. To this end, this paper studies the feasibility of bus-mounted edge servers based on real traces. First, we investigate the coverage of the buses and base stations using the Shanghai bus/taxi/Telecom datasets, which shows a great potential of bus-based edge servers as they cover a great portion of geographic area and demand points. Next, we build a mathematical model and design a simple greedy heuristic algorithm to select a limited number of buses that maximizes the coverage of demand points, i.e., with a limited purchase budget. We perform trace-driven simulations to verify the performance of the proposed bus selection algorithm. The results show that our approach effectively handles the dynamic user demand under realistic constraints such as server capacity and purchase quantity. Thus, we claim: bus-mounted edge servers for vehicular networks in urban areas are feasible, beneficial, and valuable.

💡 Research Summary

The paper investigates the feasibility of deploying edge computing servers on public buses in urban vehicular networks, using extensive real‑world trace data from Shanghai. The authors begin by highlighting the limitations of fixed‑site edge servers (e.g., roadside units or base stations), which have static locations and capacities that cannot adapt to the highly spatio‑temporally varying demand of vehicles. While unmanned aerial vehicles have been proposed as mobile edge nodes, their payload and battery constraints limit practical deployment. Buses, by contrast, offer ample space, stable power supply, and a mobility pattern that closely follows traffic hotspots, making them attractive candidates for edge hosting.

Three large datasets are employed: (1) a Telecom dataset containing over 7.2 million records from 9,481 mobile phones across 3,233 base stations, (2) a Bus dataset with more than 5 million GPS points from 2,100 buses (February 2007), and (3) a Taxi dataset with roughly 85 million records from 11,936 taxis (April 2018). The study area is a 72 × 72 grid (each cell 200 m × 200 m) covering the central part of Shanghai, yielding 934 base stations, 2,143 buses, and over 26 million taxi points. Time is discretized into 30‑second slots, allowing the authors to evaluate coverage at fine granularity.

Coverage is measured using three metrics: (a) Area Coverage – the proportion of grid cells within the service radius of any server; (b) Point Coverage – the fraction of demand points (taxi GPS records) that lie within a server’s range assuming unlimited processing capacity; and (c) Point Coverage with Capacity – the same as (b) but respecting a per‑server request limit C_bus. Experiments show that with a service radius of 0.8 km, 500 randomly selected base stations achieve >90 % area coverage, but when a realistic capacity limit of 50 requests per slot is imposed, coverage drops sharply, indicating that capacity, not range, is the bottleneck. For buses, even with 900 randomly chosen vehicles and the same radius, about 40 % of demand points can be served under the capacity‑limited metric, demonstrating that buses naturally traverse high‑traffic zones and can complement fixed infrastructure.

To formalize the selection problem, the city is modeled as a set of grids G and time slots T. Each bus trajectory r_j provides a location per time slot, and each bus i has a binary selection variable s_i (1 if equipped with an edge server). Demand points at time t are denoted P(t). The objective is to choose exactly K buses (budget constraint) that maximize the total number of demand points served across all slots, subject to (i) per‑bus capacity C_bus, (ii) service radius R, and (iii) binary decision variables. This yields an integer quadratic programming formulation, which is NP‑hard.

Given the computational difficulty, the authors propose a simple greedy heuristic: for each iteration, compute for every unselected bus the number of demand points it could serve alone (closest‑first assignment until capacity is reached). Select the bus with the highest marginal coverage, remove the covered points from the demand set, and repeat until K buses are chosen or all points are covered. The algorithm runs in O(K · M · T) time and is straightforward to implement.

Simulation results compare the greedy method against a random‑selection baseline (no base stations are assumed to be deployed). As K increases, the greedy algorithm’s coverage grows nearly linearly, outperforming random selection by a large margin. Sensitivity analysis varying C_bus (8, 20, 50) shows that the algorithm adapts well to different capacity constraints, still achieving substantial coverage. The study confirms that bus‑mounted edge servers can meaningfully alleviate the capacity shortfall of fixed RSUs, especially during peak traffic periods.

The paper concludes that bus‑based edge computing is both feasible and beneficial in dense urban environments. It acknowledges limitations such as ignoring power consumption, cooling, handover latency, and the stochastic nature of bus schedules. Future work is suggested in areas like multi‑service resource scheduling, dynamic bus selection driven by real‑time traffic prediction, and integration with existing fixed edge infrastructure. Overall, the work provides a data‑driven validation of mobile edge nodes on public transport and offers a practical heuristic for budget‑constrained deployment.