Always-On LTE Roaming

I've been thinking a bit about LTE and international roaming lately and just realized that mobile network operators need to come up with a new billing scheme compared to current systems. Here's why:

2G and 3G devices only request the establishment of a data bearer (a PDP context in 3GPP talk, or getting an IP address in Internet talk) when an application requests it for the first time. Thus, from a billing point of view, nothing is charged until that point. With LTE, however, the device gets an IP address right when the device registers with the network after startup. In effect, the 2G / 3G packet call becomes history with LTE. While in the home network this can probably be managed quite well from a billing point of view, I wonder how network operators will proceed for roaming. After all, most users will probably not be too happy to be charged just for switching on their device.

For LTE USB dongles, this might not be a problem as the user can decide whether to plug it in or not. For notebooks with a built-in LTE modem, however, or an LTE capable smartphone, things are different. A user of an LTE capable smartphone probably wants to use it abroad as well, even if it is only for voice calls and the offline organizer functionalities without being charged if he doesn't actively use the Internet. I wonder how this will be solved in practice!?

I could imagine several solutions:

  • The device detects the roaming scenario and asks the user whether to attach to LTE and get an IP address and warns the user that this might be a chargeable event.
  • The device detects the roaming scenario and doesn't attach to the LTE network. Instead, a 2G or 3G network is selected where getting an IP address right away is not required. The question then is how the user could trigger this later-on. In case of a smartphone it could wait till an application tries to access the Internet and then reselect to LTE once the connection is established. That won't work for LTE capable notbooks, though, as there are always applications crying for IP connectivity…
  • The home network detects that the user is roaming and blocks initial access to the Internet. Then, via a web based landing page, the network informs the user that different rates will apply if he proceeds. The problem with this approach is that the user has to open the web browser first before his other applications can get access to the Internet.
  • A certain amount of data traffic while roaming  is already included in the subscription. When going beyond this amount, access is blocked until the user is informed (e.g. via SMS or a landing page) that further Internet access will be billed separately and the user has given his consent.

Hm, it all doesn't sound convincing yet. Better ideas, anyone?

6 thoughts on “Always-On LTE Roaming”

  1. Roaming and LTE, hmmm, I think your option 3 breaks the user experience (not getting voice calls? but my phone is on…). Option 1 seems the most straightforward for the poor user but will require device support and standardisation.

    LTE voice roaming, specifically, still seems rife with problems. On the one hand you need local break out (like GSM/UMTS today) to lower latency, but then potentially your home HSS will route everything back there anyway for other services. And if you don’t break out locally then you can have quite long latency just on the interworking bit, esp. if you’re using a GRX/IPX (example, NZ->Singapore[IPX]->Netherlands[IPX]->Germany). Emergency calls still need local breakout, and dialling local dialplan numbers (freecall 800’s etc) becomes problematic.

    Nothing insoluble, but the additional complexity and device requirements are annoying. Even now roaming has some of these issues with home-network free and premium content (pay per event/click) needing to be hidden or otherwise presented differently for roamers. The story of the “free TV” [streamed video] watching businessman getting a huge bill when he travelled is, I believe, a common one.

    I don’t know if GSMA has recommended IMS or LTE roaming yet?

    Billing…yes. The foreign network only sees data for non-local breakout. So how can they make money charging for a non-differentiated service? Charging each data session as though it were carrying voice is not sustainable, so roaming voice calls get massively cheaper unless DPI is used, or QoS is used for billing. And then phones will just use best-effort QoS, which as we know works quite well, and/or tunnel 🙂

    Will carriers implement all this additional technology and interworking only to undermine their roaming revenue (similar to SMS displacement by IMS IM)?

  2. The presumption is that the roaming billing is based on duration and not on quantity. Would charging per transferred bit (tx and rx) avoid this?

  3. Hi Lonnie,

    I was actually assuming that future billing would be based on quantity like almost all operators do already today.

    The question is how much is the user charged even for small quantities of data which would be required for signaling and general background noise. And then how could that be distinguished from instant messaging for example which also only needs very little data.


  4. As far as I know most operators haven’t thought much about roaming situations yet — they’ve got enough to worry about rolling out their own LTE networks. Pop-up windows with message warning about roaming is generally not a good user experience, because the consumers are already used to the current roaming scenarios – you got off the plane and turn on the radio – it just works right away. And you roughly have a good idea what to expect from a billing point of view. From a NW point of view the operating bands certainly don’t help either. Even today many big operators don’t have roaming agreements for 3G data. There’re other technical issues as well…e.g. if a European user comes to Canada – he wouldn’t be able to do video steaming from youtube because the local 3G operator here bans RTSP.

    I would expect that for a long time after initiall LTE roll out roaming would fall back to EDGE or UMTS. It’s safer and easier to control.

  5. Hi there,

    not so sure that RTSP wouldn’t work for roamers in Canada. If we leave aside the pricing issue for a second, I think the blocking is either done on their GGSN or on a node behind it. Roamers, however, would use the GGSN in their home network by default. Their traffic would thus not be inspected or blocked by the Canadian operator but would be subject to what happens in the home network.


Comments are closed.