In industrial internet deployments, the industrial router serves as the core hub connecting devices to the network. Selection decisions directly impact system stability, response speed, and long-term operational costs. However, when choosing between 4G and 5G industrial routers, many users focus excessively on the single parameter of "device capacity" while overlooking the scenario-dependent logic behind it. This article systematically examines the decision-making process across four dimensions: technical differences, the real meaning of device capacity, selection methodology, and future trends.
The essential difference between 4G and 5G is not simply "fast versus faster," but a systematic expansion of network capability boundaries.
Bandwidth: From "Sufficient" to "Ample"
Under the 4G LTE Cat4 standard, the theoretical peak rate is 150 Mbps. However, in real industrial environments — affected by signal attenuation, electromagnetic interference, and concurrent users — effective bandwidth is often below 100 Mbps. This is more than enough for single-sensor data reporting. But when facing multiple HD video backhauls, LiDAR point cloud data, or large-scale equipment status collection, uplink bandwidth becomes the bottleneck.
5G, through technologies like millimeter wave, Massive MIMO, and carrier aggregation, increases peak bandwidth to 1-10 Gbps. This means that even in complex industrial sites, multiple 4K industrial cameras, AR remote guidance systems, and massive numbers of sensors can transmit data concurrently.
Latency: From "Second-Level Response" to "Millisecond-Level Control"
Industrial control is far more sensitive to latency than consumer applications. End-to-end latency on 4G networks typically ranges from 50 to 100 milliseconds. For non-real-time tasks like environmental monitoring or inventory tracking, this is acceptable. But for scenarios such as AGV path coordination, robotic arm synchronization, or PLC closed-loop control, 50 milliseconds of delay means the device state has already changed by the time the command arrives.
5G, through edge computing (pushing processing capability to the network edge) and network slicing (allocating independent virtual channels for different services), compresses end-to-end latency to 1-10 milliseconds and provides 99.999% reliability guarantees. This makes "wireless replacing wired" a reality in industrial control.
"Device capacity" is the most frequently cited parameter in router selection, yet it is also the most easily misunderstood. It is not a fixed number but a dynamic indicator closely tied to device type, data profile, and concurrency strategy.
A telling case study: In one AGV manufacturing facility, over 60 AGVs were initially connected via a 4G router. Operation was stable at first. But as the number of AGVs grew and the dispatch algorithm was upgraded, network congestion caused dispatch latency to increase by 30% and raised the risk of vehicle collisions. After switching to a 5G router, a single device stably supported more than 200 AGVs, with latency dropping below 5 milliseconds.
This case reveals three layers of meaning behind device capacity:
Device behavior determines capacity: A single camera continuously streaming HD video consumes network resources equivalent to dozens of sensors that only periodically report temperature values. The upper limit of device capacity depends on the number and characteristics of the "heaviest" devices.
Concurrency rate is a critical variable: Not all devices communicate at peak rates simultaneously. If device data reporting is staggered, a router can effectively support far more devices than its theoretical limit. Conversely, if many devices need synchronous real-time communication (e.g., multiple AGVs reporting positions simultaneously), the demands on the router's processing power rise sharply.
Uplink/downlink asymmetry: Industrial scenarios typically involve large volumes of data flowing from devices to the cloud (uplink), which differs from the download-dominated pattern of home broadband. Selection requires attention to the router's uplink processing capability and buffering mechanisms.
Thus, discussing device capacity without context is meaningless. A sound approach is to survey the site's peak concurrency rate, average throughput per device, and maximum allowable latency, then add a 30-50% margin for selection.
In real projects, a layered decision model is recommended, avoiding simple either-or judgments.
Dimension 1: Budget Constraints
5G industrial routers currently have higher hardware costs than 4G models (approximately 40-60% more). Additionally, 5G private network construction or carrier data plan expenses must be considered. For small and medium enterprises with limited budgets, a hybrid deployment strategy is advisable: deploy 5G in core production areas (e.g., automated lines, AGV dispatch zones) while continuing to use 4G in supporting areas (e.g., warehouse environmental monitoring, non-real-time data collection).
Dimension 2: Requirement Layering
Classify devices to be connected into three tiers based on network requirements:
| Device Tier | Typical Devices | Network Requirement | Recommended Technology |
|---|---|---|---|
| Critical Real-Time | AGVs, robotic arms, PLCs | Low latency (<10ms), high reliability | 5G |
| Ordinary Data | Sensors, meters, thermometers | Moderate bandwidth, non-real-time | 4G |
| High Bandwidth | HD cameras, LiDAR | High uplink bandwidth, stable streaming | 5G |
Based on the number of devices in each tier, estimate total bandwidth and connection requirements, then allocate across specific router quantities and positions.
Dimension 3: Site Adaptation
5G signals attenuate more significantly than 4G when penetrating obstacles like metal and concrete. In dense workshops, basements, elevator shafts, and similar spaces, 5G coverage quality must be assessed. When necessary, indoor distribution systems (e.g., pico cells, micro cells) or 5G private network solutions can be deployed. 4G retains advantages in wide-area coverage and penetration, making it particularly suitable for open yards, cross-facility areas, or long-distance scenarios.
In actual operations, some users attempt to run routers beyond their stated device capacity — by expanding the DHCP address pool or adjusting connection limits to connect more devices than specified. Technically, temporary, light overloads may work. However, long-term or high-load overcapacity introduces three major risks:
Sharp performance degradation: When device capacity exceeds the design threshold by 30%, router CPU utilization often exceeds 80%, leading to increased packet loss, higher retransmission rates, and lower actual throughput.
Stability and security risks: A large number of devices competing for IP addresses and connection resources can cause IP conflicts and connection resets. Simultaneously, high device density expands the network attack surface, and logging systems may become overloaded, missing security events.
Reduced hardware lifespan: Sustained overload accelerates aging of chips and power modules. Data suggests that continuously overloaded routers may see mean time between failures (MTBF) reduced by over 60%.
Therefore, adequate device capacity margin should be reserved during selection rather than attempting to challenge hardware limits.
Looking ahead 3-5 years, the industrial internet networking landscape will be characterized by 5G leading, 4G complementing — not replacement but layered usage based on scenarios.
5G's core strongholds:
Real-time motion control in smart factories
AGV fleet dispatch in logistics parks
Remote operation in ports
Unmanned operations in mines
Machine vision quality inspection requiring high uplink bandwidth
4G's enduring value:
Remote wide-area scenarios: Oil fields, power transmission lines, remote weather stations where 5G coverage is still incomplete — 4G networks are mature and reliable.
Low-power, sparse connections: The many sensors that only periodically report small amounts of data — 4G remains highly cost-effective.
Temporary or mobile projects: Construction sites, exhibitions, temporary command posts — 4G's plug-and-play convenience and low cost offer advantages.
Additionally, as RedCap (lightweight 5G) technology matures, "mid-speed 5G" solutions with cost and power consumption between 4G and traditional 5G will emerge, further enriching the selection spectrum.
Industrial router selection is fundamentally a matching of technical capabilities with business requirements. Device capacity is an important reference metric, but it gains meaning only within a specific context. Users need to ask themselves not "which is better, 4G or 5G?" but:
How many devices are on site? Is their data profile bursty or continuous?
What is the latency tolerance? Milliseconds or seconds?
Where is the uplink bandwidth pressure point? Does it need to support HD video or point cloud data?
How does the site environment affect wireless signals? Is indoor coverage supplementation needed?
Once these questions are answered, the selection becomes clear. In a landscape where 4G and 5G will coexist for years, the most rational approach is layered, hybrid deployment — invest in 5G where performance gains matter most, and continue with 4G in ordinary scenarios to control costs, ensuring every dollar of network investment delivers maximum benefit