One thing that astounded me when reading the 4G specs for the first time many years ago was that there was no separation anymore between attaching (registering) with the network after power-on and activation of a default bearer (getting an IP address). It looks like this will change again in the 5G core network.
From a mobile network point of view combining the two procedures was novel. In 2G and 3G networks, these processes were separate and especially in the early days it was not uncommon for a mobile to register to the 3G network but not request an IP address right away. These were the days when mobile networks were still circuit switched voice and SMS focused. Only a few years later with the rise of Android and the iPhone did it become natural for devices to perform both procedures at the same time as Internet connectivity became the norm rather than a fancy exception.
From an ‘Internet’ perspective it also seems pretty normal that connecting to the network also implies getting an IP address straight away. After all, there is little use to connect to a network if you can’t communicate in it. 4G LTE specs also went along this line of thinking which created quite a bit of confusion and trouble up to present days as many people got used to thinking that being registered and having a context (IP address) are two different concepts. Event today some people still talk about ‘packet calls’ when they talk about ‘Internet connectivity’.
So to my great surprise I noticed that in the 5G core specification, the registration procedure and the session establishment procedure are again two separated procedures! To compare, have a look at figure 5.3.2.1-1 in TS 23.401 where the LTE attach and create session procedures are clearly combined into a single process. In 5G, however, they are completely separate and are even separated in the specification document by dozens of pages. Registration is described in 3GPP TS 23.502, Chapter 4.2.2 while creating a PDU session (the same as establishing a Default Bearer in LTE, i.e. getting an IP address) is described in Chapter 4.3.2.
So the way I read this is that in 5G it is possible again to be registered to the network but not to have an IP address. Hm, I wonder why the concept was changed again, did a ghost of the past attend a 3GPP meeting?
Does the situation with 4G means that ‘disable mobile data’ mode in Android is a fiction?
Well, that depends on how you look at it. Android will close the virtual network interface that provides Internet connectivity. In the modem chip, however, the connection (the default bearer) and the IP address lives on. When the user enables mobile data again, the virtual network interface in Android is connected to the modem again. If you look at this from a network point of view: When the user disables mobile data, nothing changes on the network side, it doesn’t notice, it’s not informed. Nevertheless, “disable mobile data” does what it is supposed to do, it makes sure apps can’t communicate with the Internet and don’t eat up your monthly data bucket while data is disabled.
“Disable mobile data” may not totally prevent usage of mobile data.
Some devices do allow “IMS registration” (for VoLTE & WiFi Calling) even when mobile data is disabled.
I have also come across many cases where users disabled “Data Roaming” but were still charged for some small amount of usage (typically less than 1MByte).
(am working for a mobile telco’s cust care team)
Isn’t this logical result of the new services ethernet and unstructured (non-IP) which don’t require an IP address?