UMTS Femto cell solutions are being announced lately by both startups and established players such as Nokia Siemens Networks. Leaving aside the question of whether femto’s make sense or not I tried to find out how femto’s can be integrated with the macro layer of the network. It seems not to much information is available about the technical part on the net. So here are my questions, maybe some of you know more. If so please consider leaving a comment.
Basically I’ve seen two approaches to femto. For both cases, the cells are connected to the network via DSL or cable:
Pure Base Station Approach
In this approach the femto cell is included as part of the overall radio network. This should require configuration of both the femto cell and the macro cell layer for handovers and cell reselection. I’ve seen some patent applications from Ericsson which describe that the pico cell is equipped with a receiver that can scan the environment for neighboring cells. The result is then reported to the network which in turn sends the required neighboring cell lists to the femto for broadcast. Nothing is mentioned, however, of how the macro layer is configured. If this is not done, I wonder how a mobile in idle mode can change to the femto cell.
Also, I wonder if it is foreseen to restrict access to a femto cell to the owner of the cell? After all, if I had a femto cell at home, would I want my 25 neighbors to also use it for free? If access can be restricted how is that done? Should the macro layer broadcast the cell info for my femto cell, others will see it as well. And if it doesn’t, how can my own mobile detect the femto cell once I arrive home and still have coverage from the outdoor macro cell?
Speculation: While a mobile has a connection established to the network it can be asked to report cells which are not in the cell info list (the so called "detected cells" broadcasting on the same frequency as the current cell). This could then be used by the RNC in combination with with my user ID to decide whether to hand over the connection to the femto cell, in case the owner of the mobile is the owner of the femto or to leave the call on the macro layer. This does not work when the mobile looses connection to the macro cell layer, however. In this case the network search of the mobile will detect the femto and the mobile will try to attach. How can this be gracefully prevented in case the femto only allows a select few users?
Scalability: If it can be avoided that the macro cell layer has to broadcast information about the femto cell layer then I don’t see scalability issues on the Node-B side. If it can’t be avoided then I wonder how the solution scales. A macro layer cell is usually designed for about 2000 users. If ‘only’ 50 of them use a femto cell at home I wonder how this can be accommodated for in the neighboring cell list!?
The Network In A Bottle Approach
Another femto approach used for example by 3WayNetworks is to combine the complete functionality of the network from base station to MSC into the femto base station. On their web site, 3WayNetworks mentions that the femto base station can use a different Mobile Country Code and Mobile Network Code and thus runs completely independent from the macro layer. This might make rejecting unauthorized users a bit simpler than in the approach above but still leaves open the question of how authorized mobiles find and use the cell in the first place in case the macro layer is still strong enough where the femto cell is to be used (e.g. to increase overall network bandwidth).
Speculation: Here, an old GSM trick could help which probably still exists for UMTS: For national roaming the mobile can be instructed to scan for the home network every couple of minutes. Femto subscribers could be given a SIM card which the femto’s MCC/MNC as home network. Thus, femto subscriber mobiles would keep looking for femto cells while other subscribers could automatically be barred. UMTS also knows the concept of equivalent network which might also help here (see 3GPP TS 22.011 chapter 3.2.2.5).
A lot of questions… If you have an answer, please leave a comment.
another option would be: Femto-over-VPN-over-commodity broadband
i feel this would be nice for providers that have both CMTS and broadband infrastructure. therefore, they can take their little bit of CMTS network to any place with a broadband connection.
Hello hip2b2,
thanks for your comment, you are right! I implicitly assumed that femtos, no matter which approach is used, are connected to the network via DSL or cable. I’ve updated the text above a bit to make this explicit.
Thanks again!
Martin
Using a different MCC and MNC does not seem to be a very clean solution. I am sure the standards folks would like to shy away from that.
As for the first option, it does seem to be a problem I have never given thought of before. Why wouldn’t I want to share my femto? Maybe, a carrier can give incentives to Femto-owners to encourage Femto sharing?
If we really did want to restrict access … maybe a femto has a white list of allowable IMSI/IMEI pairs?
There must surely be an incentive such as the one you have mentioned, otherwise people will wonder why they should share.
White list in the Femto won’t work because how can you make mobiles stay away from the cell? Not so easy – But sharing is better anyway – if the incentive for it for the end user is right.
Cheers,
Martin
Different femtocell vendors have come up with different technical solutions for these issues.
At ip.access, we’ve solved Access Control by storing an allowed IMSI list locally at the femtocell, so that unauthorised mobiles will be rejected without sending signalling traffic to the core network. These unauthorised mobiles will remain on the macro network. End users simply add allowed phone numbers (MSISDN, not IMSI) on a secure web page.
Most operators want Access Control implemented. This puts the control in the hands of the consumer. The feeling is that consumers will want to avoid the risk of neighbours hogging their internet bandwidth (e.g. by watching YouTube on their phones all day).
Handover is another tricky issue for which we’ve had to develop novel technology. Hand-in from the macro network to a femtocell is particularly difficult, because the macro cell can have potentially hundreds of femtocell neighbours sitting inside it and it can’t handle such a big neighbour list. We’ve patented a solution for this.