Definition
Location Update (Location Update) is the process by which the terminal reports its current location area (LA/RA) or tracking area (TA) to the network. It is a core process of mobility management. When the terminal enters a new TA/LA, triggers periodically, or powers on to attach, it performs a location update to inform core network elements like MME/AMF of the terminal's location, enabling correct paging. In 4G it is called TAU (Tracking Area Update), in 5G it is Registration Update.
Trigger Conditions
- Entering new TA/LA: The terminal measures and finds the current TA is not in the stored TA list, initiates TAU
- Periodic update: Timer expires (e.g., after 54 minutes) to prevent the network-side state from expiring for long-stationary devices
- Power on/Attach: When the terminal powers on or recovers from offline, it needs to perform attach or registration, including reporting location information
- Cross MME/AMF: When service nodes change due to roaming or load balancing
Processes and Optimization
TAU Process: Terminal initiates TAU Request → MME validates and updates location in HSS → Returns TAU Accept (including new TA list). To reduce signaling, the network can configure multiple TAs as a TA List; movement within the list does not trigger a TAU. IoT devices can adopt an extended periodic TAU (up to about 310 hours), reducing power consumption and signaling overhead.
Typical Use Cases
Location update collaborates with paging, handover , roaming to ensure service reachability. SimtoLink IoT SIM cards automatically complete TAU in the operator network, and the cloud platform can display device TA information, providing connectivity assurance for remote wake-up and downlink control.
Location update enables the network to correctly page IoT devices, which is key for mobility management. SimtoLink IoT SIM cards automatically complete TAU and other location update procedures in the operator network. The SimtoLink cloud platform can display device location area information, ensuring connectivity availability for IoT scenarios requiring remote wake-up and downlink control.
Signaling Differences Between Border and Periodic Updates
A location update exists so the area stored in the core network still covers the cell the device is in. Otherwise downlink data has no way to decide which group of base stations should page it. A border update happens when the device finds that the current tracking area is not in the list it saved. A periodic update has nothing to do with movement: when the timer expires, the device reports once, so a device that has been still for a long time is not silently removed on the network side. Registration at power-on, or when the device returns from power-off or from being off the network, also carries the current location, but that builds a context. It is not only a location refresh.
Both the 4G TAU and the 5G registration update do this job. The signaling lands on the MME or the AMF, not on the user-plane gateway. So when a location update fails, the user-plane IP may still be there while paging is already useless. This is different from handover. Handover changes the serving cell while the device is connected. A location update mainly happens while the device is idle, or when it has to register again. Mixing the two counters makes ordinary movement look like a mobility fault.
How the List, the Timer, and Power-Saving Mode Fit Together
Only when the network sends a list of several tracking areas can the device move among those areas without updating every time. The network decides the list. The device cannot enlarge it on its own. The periodic timer is also sent by the network in the accept message. If an IoT device negotiates PSM or a longer period, the value that actually applies is the stricter of what the device requested and what the network allowed. If the network has not enabled the feature, writing a very long period on the device does nothing.
The settings have to match how traffic arrives. Stretch the period only when the device almost always sends uplink and can tolerate being unreachable for a while. If the platform must deliver a command within a few minutes, the period and the paging watch in the RRC state machine have to be much shorter. Across operators, or across core-network pools, also confirm that the new node can fetch the subscription. Otherwise the update is rejected, the device retries over and over, and it uses more power than ordinary movement. A slow device wandering on a cell reselection boundary will burst updates if neighboring cells sit inside and outside the list.
A Gap in the List Leaves a Paging Hole
The easy mistake is to read “the device is online” as “its location is still valid.” If the periodic timer is shorter than the traffic interval, the device wakes often just to refresh its location. If the timer is longer than the network’s implicit detach time, the core has already cleared the context while the platform still sends to the old address, and the device is unreachable now and then. Another case is a list plan that slices a highway or a campus boundary too finely. A vehicle updates at every junction, signaling congests, and then the update itself fails.
A field mistake is to use the administrative district in the base-station planning data in place of the tracking area. District borders and tracking-area borders do not coincide. A device can cross an area without leaving the city. Check the list stored on the device, the tracking area of the current cell, and the time the core last accepted an update. Signal strength alone does not explain it. When an update is rejected, read the cause: “tracking area not allowed,” “congestion,” and “illegal device” need three different responses.
After a periodic update is rejected, the device backs off and tries again as specified. If the platform keeps sending during that backoff, the failures pile up and registration does not get faster. An update across pools may also pull subscription data. A fetch timeout and a radio failure have different delay patterns: the first clusters around a fixed signaling round trip, the second moves with the signal level. Plot both on one drop curve and the optimization goes the wrong way. If a stationary device is pulled by neighbor-cell breathing, it reselects and updates every few minutes. The current trace then shows a spike on a fixed beat. That spike is the location procedure itself, not an application report.