I got a question long time ago, saying "Why we see measurement control/measurement report almost always in Handover procedure ?" The short answer is that the two are separate procedures, and the network links them on purpose.

This page first answers that question with a simple two-cell example. It then shows what the MEASUREMENT CONTROL message carries, and which measurement event usually leads to which handover procedure in 25.331.

• Why does Measurement Control appear in almost every handover ?
• How does the network set up a measurement ?
• Which measurement event leads to which handover ?
• Reference

Why does Measurement Control appear in almost every handover ?

In real live network, it would be common to perform 'Measurement Control' both for SHO and HHO. But purely in terms of technology, 'Measurement Control' and 'Measurement Report' is not a mandatory part for any Handover process. If I explain these two procedure in a very simplified (a little bit over simplified) manner, I would say

i) Handover - Changing a communication cell from one to another, or changing the number of communication cell to one to another.

ii) Measurement Control - Ordering UE to perform a specific measurement (e.g, the signal quality of the current cell or the signal quality of another cell).

Basically these two procedures are independent.

But in reality, most of network operator combine these two process mainly to improve the successbility of the handover.

For example, let's suppose that there are two cells (let's call them Cell A, Cell B respectively) and a UE is now communicating to cell A and now your network is trying to let UE change the communication cell from Cell A to Cell B. The simplest way would be to send UE a handover command, saying "change Cell from A to B". But what if the cell B is not ready to communicate with your UE. For example, the signal quality between Cell B and UE is so bad and it will almost for sure that the call would drop as soon as the UE changes the cell ?

Do you still want to change the cell ?

The answer would be No.

Main for this kind of reason, the network try to make it sure that the target cell (the UE is handed over to) is good enough condition before trying handover. That's why in most live network log, you would see measurement control/measurement report message before the handover.

A handover without a preceding measurement is still possible, and engineers call it a blind handover. The RNC then relies on its own planning data, for example a co-located cell on another frequency with the same coverage. It saves the time of a measurement, and on an inter-frequency or inter-RAT target it also saves the compressed mode gaps. But the RNC then has no proof that the UE can hear the target cell, so the risk of a failed handover is higher.

• The two procedures are independent in 25.331 : No handover procedure requires a MEASUREMENT REPORT before it.
• Operators link them to protect the call : The report proves that the UE can receive the target cell well enough.
• A blind handover skips the measurement : It is faster, but it trades that speed for a higher failure risk.

How does the network set up a measurement ?

Let's look at what the UE actually receives. In CELL_DCH, the RNC sends a MEASUREMENT CONTROL message on the downlink DCCH with AM RLC, and one message sets up, modifies or releases one measurement. The UE answers later with a MEASUREMENT REPORT on the DCCH, using AM or UM RLC as the message asked.

The main IEs of the message in 25.331 v19.0.1 clause 10.2.17 are these.

• Measurement Identity : A number from 1 to 16, or up to 32 with the -r9 type. The UE puts it in every report, so the RNC knows which measurement the report belongs to. The RNC may use several identities for the same measurement type.
• Measurement Command : setup, modify or release.
• Measurement Reporting Mode : The transfer mode, acknowledgedModeRLC or unacknowledgedModeRLC, and the trigger, periodical or eventTrigger.
• CHOICE Measurement type : Intra-frequency, inter-frequency, inter-RAT, UE positioning, traffic volume, quality and UE internal measurement. Later releases add CSG proximity detection, E-UTRA measurement for CELL_FACH and application layer measurement configuration.
• Additional measurements list : Other measurement identities whose results the UE adds to each report of this measurement.
• DPCH compressed mode status info : This activates or deactivates the gap pattern that an inter-frequency or inter-RAT measurement needs.

Figure 1 shows the most common chain in a live log, a soft handover driven by event 1a. The RNC first sets up an intra-frequency measurement with event triggered reporting. When a cell of the monitored set enters the reporting range, the UE sends a MEASUREMENT REPORT for event 1a. The RNC then adds the radio link of that cell with ACTIVE SET UPDATE.

Figure 1. Measurement control followed by a soft handover. The handover message comes only after the report, so the RNC adds Cell B only when the UE has measured it in the reporting range.

For an inter-frequency or inter-RAT measurement, the chain is longer. A UE that cannot measure another frequency while it receives the serving one needs compressed mode gaps. The gap pattern itself comes in a message such as PHYSICAL CHANNEL RECONFIGURATION, and the MEASUREMENT CONTROL can activate it through DPCH compressed mode status info. The Compressed Mode page shows such a configuration from a live log.

• One message, one measurement : Each MEASUREMENT CONTROL sets up, modifies or releases the measurement with one Measurement Identity.
• The Measurement Identity links report to request : Read it first when you match reports to measurements in a log.
• Gaps come with the measurement : Inter-frequency and inter-RAT measurements usually need compressed mode, and the message can switch the pattern on.

Which measurement event leads to which handover ?

Once the UE reports, the RNC decides on the handover type. 25.331 does not tie an event to a handover procedure, so the mapping below is the usual operator practice, not a rule. Still, each event family fits one handover type well, because each family watches a different kind of target.

The table lists only the events that commonly drive a handover. Events 1e, 1f and 1j, 2a, 2c, 2e and 2g, and 3b and 3d exist as well, and the Measurement Event page explains the thresholds and the reporting range. A hard handover with change of frequency also has a side effect in 25.331 clause 8.3.5. The UE stops all intra-frequency and inter-frequency measurement reporting on the cells in CELL_INFO_LIST, and each stopped measurement restarts only when a new MEASUREMENT CONTROL arrives with the same measurement identity. So after a hard handover you will usually see a fresh set of MEASUREMENT CONTROL messages in the log.

• Event 1x usually means soft handover : The follow-up is an ACTIVE SET UPDATE, not a hard handover.
• Event 2x and 3x usually come with compressed mode : The UE needs gaps to measure the other frequency or system.
• A hard handover resets the measurements : Expect new MEASUREMENT CONTROL messages right after it.

Reference

• 3GPP TS 25.331 v19.0.1 : Radio Resource Control (RRC); Protocol specification