The Packet Call Becomes History with LTE

This week I've been taking a closer look at how LTE networks will be used in practice. One of the big differences to current 3G and 3.5G networks is that LTE is fully based on IP, i.e. there is no longer a circuit switched core network for voice telephony. Today, it makes sense for 3G/3.5G devices to attach to the network without requesting an IP address since they are (still) mostly used for voice telephony. In LTE networks, however, a device without an IP address is completely useless. Hence, the LTE network attach procedure already includes the assignment of an IP address. Consequently, an LTE device will always have an IP address. For people coming from the LAN/WLAN world this is nothing new. For people in the cellular industry, however, this is revolutionary. The 3G/3.5G  procedure of 'establishing a packet call', a term coined with the old thinking of establishing a circuit switched connection for a voice call in mind, will thus become history with 4G. Many people in the industry will have to change their picture of the mobile world to accommodate for this.

Comments

[PierreL]

Yes Martin, you got it right.
At power-on, the terminal will not only register to the network, but also get an IP, as well as a default bearer with default QoS.
The question is: does that mean a global IPv6 migration so as to solve the good old IP address shortage issue ?

[Erwan]

A lot of folks seem to be getting very enthousiast over fixed/mobile convergence, and Wifi VoIP. These are indeed the path to the future, but it is important to remind how much WLAN is lagging behind "good old GSM" in terms of RSSI, handovers, etc... those 802.11 extensions being still not widely adopted in the industry. At this point in time, GSM is still a lot more power efficient than any other technology.

[Stefan Constantinescu]

Which will come first, LTE or IPv6? Will the switch to IPv6 need to happen to support all the new LTE devices on the market and thus we can finally stop it with the multilayer NATs?

[Martin]

I'd put my bets on LTE coming long before IPv6 gains wider popularity. I haven't really seen IPv6 moving ahead on the application layer and in live networks in the past years and see no reason why it should do so in the next years (despite lack of IPv4, etc.). I wonder what kind of big bang has to happen before it finally starts moving in earnest.