Why Can Devices Stay Connected After Attach? — The Mechanics Behind Handover

Introduction

When connecting IoT devices to a cellular network, a natural question arises:

“Why does communication continue even while the device is moving?”

In this article, we break down that behavior from the perspective of handover.

---

State Before Handover

In the previous article, “Why Can Devices Access the Internet After Attach?,” we explained how the GTP tunnel is established. At this stage, the GTP tunnel has already been established, so Internet connectivity is available through the path:

“UE → Serving eNodeB → S-GW → P-GW → Internet”

Here, we also need to touch on RRC Connected state and measurement control and RRC Connection Reconfiguration as additional prerequisites.

RRC Connected State and Measurement Control

After Attach, the UE enters the RRC Connected state and maintains a control-plane connection with the Serving eNodeB. In this state, the UE does not simply communicate; it also performs radio measurements based on conditions instructed by the eNodeB.

What Is RRC Connection Reconfiguration?

The conditions used to determine whether handover should be executed are configured in the UE through RRC Connection Reconfiguration. These parameters include:

• Measurement targets (neighbor cells / frequencies)
• Measurement events (for example, Event A3)
• Thresholds (offset)
• Time-to-Trigger (only considered valid if the condition continues for a certain period)

Based on these conditions, the UE compares the radio quality of the Serving Cell and Neighbor Cell (RSRP / RSRQ), and sends a Measurement Report only when the configured conditions are satisfied.

👉 In other words, the handover trigger does not occur through an autonomous decision by the UE. It occurs according to rules predefined by the eNodeB.

---

Handover Process 1 (First Half)

Let us first look at the first half of the handover process.
At this stage, the path has not yet been switched, and forwarding occurs in the form:

“UE → Target eNodeB → Serving eNodeB → S-GW → P-GW → Internet”

※ At this point, the S-GW is still sending data to the Serving eNodeB. Therefore, the Target eNodeB receives data via the Serving eNodeB.

---

Handover Process 2 (Second Half)

Next, let us look at the second half of the handover process. At this point, the path:

“UE → Target eNodeB → Serving eNodeB → S-GW → P-GW → Internet”

is fully switched to:

“UE → Target eNodeB (the new Serving eNodeB) → S-GW → P-GW → Internet”

👉 This is achieved by the Path Switch, which updates the forwarding destination of the S-GW to the Target eNodeB.

---

Roaming During High-Speed Movement

During high-speed movement, cell crossing events occur continuously.
Depending on the environment and cell size, it is not unusual for handovers to be repeated every few seconds.

The key point here is that handover, which is the core process enabling communication continuity, is executed extremely quickly.

As discussed in this article, the entire sequence included in a single handover cycle — from trigger condition satisfaction to Path Switch completion — is completed on the order of milliseconds.

Cell size varies depending on base station deployment, but is generally formed on the following scale:

• Macro cells: several hundred meters
• Small cells: several tens of meters

Now let us consider an extreme case in which the UE crosses cells every 10 meters.

In that case, the relationship between handover processing time and movement speed is as follows:

---

Conclusion

During handover, the data path temporarily becomes dual-path, and is then switched to the new route.

This is how connectivity can be maintained even while the device is moving.