For example, when the terminal device is a water meter, a main service of the terminal device is monitoring water consumption data, and reporting the water consumption data. In other words, almost all service features of the water meter are uplink trigger, and a probability that the water meter is paged for downlink transmission is relatively low. When the terminal device is a smoke alarm, a main service of the terminal device is reporting smoke data after smoke is detected. In other words, almost all service features of the smoke alarm are uplink trigger, and a probability that the smoke alarm is paged for downlink transmission is relatively low. It can be learned that the water meter and the smoke alarm have a same service feature. Therefore, the water meter and the smoke alarm may be put into a same category based on the service feature of the water meter and the service feature of the smoke alarm, and a probability that a terminal device in the category is paged is relatively low. On the contrary, when the terminal device is a mobile phone or a computer, a service feature of the terminal device includes uplink trigger and downlink paging, that is, a probability of being paged is relatively high. It can be learned that the mobile phone and the computer have the same service feature. Therefore, the mobile phone and the computer may be put into one category based on the service feature of the mobile phone and the service feature of the computer, and a probability that a terminal device in the category is paged is relatively high. In this way, a mobile phone and a notebook computer that have a relatively high probability of being paged are put into a same category, and a water meter and a smoke alarm that have a relatively low probability of being paged are put into a same group.