Understanding 5G Network Latency and the Role of MEC

In recent years, topics related to 5G have attracted great attention, ranging from autonomous driving and remote surgery to high‑profile discussions about the technology’s security. While 5G is frequently highlighted, the concrete technical advantages over 4G—especially in terms of business‑level latency—are often misunderstood.

A popular science video demonstrated a remote brain‑pacemaker implantation performed via a 5G link, mentioning an "end‑to‑end latency of 1 ms". This claim is a common misconception; the article clarifies the actual components of network latency and how 5G reduces them.

The current mobile network architectures are introduced:

4G: Traditional fourth‑generation network composed of 4G base stations (eNB) and the Evolved Packet Core (EPC).

5G‑NSA (Non‑Standalone): 5G base stations (gNB) connected to the existing 4G EPC, a cost‑effective transition solution.

5G‑SA (Standalone): Fully independent 5G core (5GC) plus 5G base stations, completely separate from 4G.

Business latency is defined as the sum of system‑processing latency and data‑transmission latency. Processing latency depends on device performance and algorithms and is generally independent of the network architecture, so the focus is on transmission latency.

In the 5G standard, the "1 ms" figure refers to the air‑interface (radio) latency between the user equipment and the gNB. For 4G, the air‑interface latency is typically around 20 ms (down to about 10 ms under good conditions). The article presents typical latency values for different network layers (illustrated in Figure 2).

Four main latency contributors are identified:

Air‑interface latency.

Transmission latency from the base station to the core network.

Core‑network processing and forwarding latency.

Internet transmission latency to the application server (approximated as 100 ms for comparison).

Using representative numbers, the end‑to‑end latency for each architecture is calculated:

4G: 10 ms (air) + 10 ms (BS‑to‑core) + 10 ms (core) + 5 ms (Internet) = 35 ms

5G‑NSA: 1 ms + 10 ms + 10 ms + 5 ms = 26 ms

5G‑SA: 1 ms + 2 ms + 5 ms = 8 ms

These figures show that, by themselves, 5G does not achieve a qualitative leap in latency compared with 4G. The breakthrough comes from integrating MEC (Multi‑access Edge Computing) with 5G‑SA.

MEC brings compute and storage resources to the network edge (e.g., at the UPF), allowing services to be deployed close to users. This dramatically reduces the transmission latency to the service, achieving sub‑10 ms end‑to‑end delays, which is sufficient for latency‑sensitive applications such as VR, cloud gaming, and autonomous driving.

The article notes that MEC can interoperate with 4G, 5G‑NSA, and 5G‑SA, but only 5G‑SA can place the service split at the UPF, eliminating core‑network processing delays. The resulting latency with MEC is:

4G + MEC: 35 ms (no significant improvement)

5G‑NSA + MEC: 26 ms (moderate improvement)

5G‑SA + MEC: 8 ms (qualitative improvement)

In summary, the combination of 5G Standalone architecture and MEC technology can reduce business latency to below 10 ms, enabling reliable low‑latency communications (uRLLC) for emerging use cases.