17 Nov 2006

VoIP Episode 2: Why do operators want VoIP/UMTS?

In my previous post, I looked at some of the reasons why mobile operators think they need VoIP/UMTS. In this post I'd like to investigate these reasons a bit further. The list is not comprehensive so if you think of any other motivations please let me know:

  1. Operators think that VoIP can help them utilise their transmission network more efficiently. The idea here is simple: if all traffic is IP based, you can carry all of it on one IP backhaul network. This may be true to a certain degree in the sense that reliance on legacy transmission networks can be slowly phased out. Then the operator eventually can consolidate the traffic on one IP network, and hence achieve cost savings due to economies of scale. Also the network becomes simpler and hence cheaper to manage and run (One box does everything instead of many boxes doing many things). However, when talking about VoIP, the amount of overhead that the network has to cope with reduces the attractiveness of any cost savings. This is the famous "IMS Tax" term that was coined to refer to the additional costs a cellular operator has to absorb to provide VoIP - here's a good example that refers to "IMS tax".
  2. They think that VoIP can help them squeeze a bit more out of their radio network. The idea is to offer VoIP calls with low rate codecs using packet channels (dedicated or shared), which has a capacity advantage over circuit switched voice. This, unfortunately, is nothing but wishful thinking. We discussed this in the previous post, so I am not going to delve into it again. In a nutshell, the IP overhead waste some of the air interface resources, and even with header compression schemes, the performance is only partially recovered.
  3. They freak out when competing providers using other access technologies (WiMax, EV-DO ..etc) are offering VoIP. UMTS providers feel the urge to support VoIP to compete. This view is rather delusional because some competitors do not have a choice but offer voice over IP. For example, EV-DO can not carry circuit switched voice (DO: stands for data only), therefore VoIP is the only route to offer voice services in this case. WiMax, is also built from the onset to carry packets and packets only, hence VoIP is again the only alternative to offer voice services. For UMTS networks, this is a non-issue really. You can carry voice much better on dedicated circuits. Now some operators may want to run their networks around the Internet model, where they see themselves as a provider of a "fat pipe" and users are free to use it the way they like. Some users may want to run VoIP clients from their terminals (e.g. Skype), and that's fair enough as long as they understand the difference in quality they'll experience. If VoIP allows me to call Venezuela, or the Gabon for a fraction of the cost of trunked calls I'll go for it even if the quality is slightly worse. I will dedicate a post in the future to discuss operators' strategy in light of the recent news about Hutchinson's 3.
  4. Convergence with their IMS networks. This was partly covered in point 1 above when we talked about all traffic types being carried over IP. "Convergence" in this case also partly refers to the additional services that can accompany voice such as presence information, contact lists, profiles ...etc . This is a valid point, and operators can perhaps create a value system around such services. Nevertheless, if operators follow the "ISP" model (provide a pipe and let the user fill it), there is no need to offer any of these services as part of the IMS network.

Please feel free to comment on my "rant" above!.

© Copyright belongs to the author. Do not cite or quote without the express permission of the author

No comments: