Content for TR 22.951 Word version: 18.0.1

A.1 Advanced Geographical split with seamless national roaming A.2 Common shared spectrum scenario

A Examples of network sharing realisations p. 13

In this annex some examples of how network sharing can be realised in real situations is described. Often the realisation of a shared network will encompass more than one scenario described in section 5. These examples aim to demonstrate how the functionalities needed identified in this technical report are applied.

A.1 Advanced Geographical split with seamless national roaming p. 13

Operators A and B, both licensed to operate a GSM network and a 3G network have agreed to share the 3G portion of the radio access network in some areas of the country where the traffic is expected to be low (Low Traffic Areas, or LTA) while building separate 3G networks in the areas with higher traffic density (High Traffic Areas or HTA). The LTA is split in two parts, one built by Operator A, the other by Operator B.

Note that a HTA may correspond to a subset of a Location Area (e.g. motorway, railway, shopping centre, train station…)

Operators A and B will not share their respective GSM networks anywhere in the country and will not share the 3G network in the HTA, but want to achieve seamless service continuity for subscribers moving among permitted networks.

Copy of original 3GPP image for 3GPP TS 22.951, Fig. A.1: mobility for a subscriber of operator A

Figure A.1: mobility for a subscriber of operator A
(⇒ copy of original 3GPP image)

Operator A and B aim to offer full mobility in the shared 3G network without loss of active services (both Circuit and Packet switched) when crossing the border between areas where coverage is provided by the other operator.

When the users exit the 3G coverage area, their UE are transferred to the 2G network of the home operator. Similarly, when the user enters an HTA coming from the home GSM network or from either of the LTA, the UE is transferred to the 3G network of the home operator.

Moreover, when the subscriber registered in the GSM network moves into the HTA it should be possible to move as soon as possible to the 3G network of the home PLMN. More generally the subscriber should move to a permitted 3G network as soon as it becomes available.

The network name displayed on the UE is that of the home operator regardless of the provider of the 3G coverage.

A.1.1 Functionalities needed p. 14

The need for the following functionalities can be identified by analysing the above architecture and mobility rules for network sharing

Both operators will want to be able to apply different mobility rules according to the identity of the subscriber (IMSI). Examples of subscribers categories are:
- subscribers of operator A
- subscribers of operator B
- inbound international roaming subscribers
  The latter category could be further divided in sub-categories
The differentiated mobility behaviour needs to be supported both for UE controlled mobility (e.g. cell reselection) and for Network controlled mobility (e.g. handover).
The differentiated mobility behaviour needs to be supported both in the UTRAN and in the GERAN.
Both operators will want the display of the operator name on the terminal to be independent of the provider of the radio coverage
Both operators will want the facility to steer the traffic to the preferred network and preferred radio access technology according to the subscriber profile, network status, service used
Both operators will want to provide services to their subscribers from their respective core networks
Both operators will want to have continuity for both Packet Switched and Circuit Switched services used while the subscriber roams from one part of the network to another.

A.2 Common shared spectrum scenario p. 14

Most of the functionalities needed for realisation of this scenario is related to scenario 4 common spectrum network sharing.

Two operators, A and B, each have their own countrywide 2G networks in operation. For 3G they agree to build a common countrywide radio access network and to connect this radio access network to the two existing, 3G enhanced, core networks. The two operators will continue with their own subscribers and compete with services and prices.

For the radio access there are three countrywide radio access networks involved: A, B, and C, where A is the 2G radio access network of operator A, B the 2G radio access network of operator B, and C the common 3G radio access network.

Each of these radio networks will broadcast their own PLMN id (PLMN Id = MCC+MNC). Let us allocate Id_A, Id_B and Id_C respectively to this parameter. Subscribers of operator A will have the MCC+MNC part of the IMSI equal to Id_A and subscribers of operator B will have the MCC+MNC part of the IMSI equal to Id_B. This means that the subscribers of operators A and B will regard the respective 2G radio access network as their respective HPLMN and radio access network C as a VPLMN. National roaming is assumed not to exist between the two operators.

There are two core networks present, namely the 2G core networks of operator A and of operator B, appropriately upgraded with the necessary 3G functionality. The nodes of the 2G radio access network A are only connected to nodes in CN_A and nodes of the 2G radio access network B are only connected to nodes in CN_B. The nodes of the 3G radio access network C are connected to both nodes in CN_A and to nodes in CN_B by some appropriately enhanced Iu interface.

The basic architecture is shown in Figure A.2. It must be assumed that all different combinations of radio coverage exist, this is also shown in the Figure.

Copy of original 3GPP image for 3GPP TS 22.951, Fig. A.2: Basic Network Sharing Architecture

Figure A.2: Basic Network Sharing Architecture
(⇒ copy of original 3GPP image)

A.2.1 Functionalities needed p. 15

Here we describe the different functionalities needed to be identified within the shared network scenario described in the Section above. The overall requirement is that it should be completely transparent to the users that operator A and operator B has chosen to share a 3G network.

Idle Mode Roaming

Subscribers of operator A

Subscribers of operator A should be served only by cells of radio networks A and C and only by core network CN_A, i.e. the service area of a subscriber of operator A is the sum of A, AB, AC, ABC, BC and C in Figure A.2. When the subscriber switches on his UE in this service area he is automatically registered in operator A's core network CN_A. The UE should display the service provider name and possibly the PLMN name according to the settings of the USIM/network/UE in the same way irrespective of the radio network used (A or C).

The possibilities are summarized in Table A.1 below.

Table A.1: Change of registered cell and network for an operator A subscriber

to cell in →	A	B	C
From cell in ↓	A	B	C
other network	Allowed. Normal registration.	Not allowed (rejection by CN).	Allowed. Registration only to CN_A.
A	Allowed. Normal location updating.	Not allowed (rejection by CN).	Allowed. Registration only to CN_A.
B	Not applicable.	Not applicable.	Not applicable.
C	Allowed. Normal registration.	Not allowed (rejection by CN).	Allowed. Normal location updating only to CN_A.

Subscribers of operator B

The handling of subscribers of operator B is exactly the same as for subscribers of operator A, with A and B interchanged.

Roaming users

There are three different types of roaming users to be considered (i) users who are allowed to roam on operator A's network, (ii) users allowed to roam on operator B's network, and (iii) users allowed to roam on both networks.

Visiting roaming users with roaming agreement with only operator A should be handled the same way as subscribers of operator A, see previous Section. Visiting roaming users with roaming agreement with only operator B should be handled the same way as subscribers of operator B

The two operators A and B may have different charging policies for roaming users. Thus one of the operators may be preferred by some roamers and the other operator preferred by other roamers. It must therefore be possible to indicate to the UE, at least when it is in radio network C, which core networks are available so that the user can perform an active selection. If such a manual selection is made, the UE should be treated the same as a subscriber of the chosen operator.
If no such selection of core network is made, some filtering is required to select one of the core networks when the UE registers via radio network C. Unnecessary change of core network should be avoided. For example, if the UE is using radio network B and CN_B and changes radio network to C, the UE should remain registered in CN_B. Only when this UE has to change to radio network A (due to radio conditions) a change to CN_A has to be made.

After a core network operator has been chosen, the UE should display the PLMN name of the CN operator currently used and possibly the service provider name according to the settings of the USIM/network/ME.

The different possibilities are summarized in Table A.2 below.

Table A.2: Change of registered cell and network for a roaming user allowed to roam on both both networks

to cell in →	A	B	C
From cell in ↓	A	B	C
other network	Allowed. Normal registration (in CN_A).	Allowed. Normal registration (in CN_B).	Allowed. Selection of core network made by the radio network or by interaction between the radio network and one or both of the core networks.
A	Allowed. Normal updating (within CN_A).	Allowed. Change to CN_B is necessary.	Allowed. Normal updating, stay in CN_A.
B	Allowed. Change to CN_A is necessary.	Allowed. Normal updating (within CN_B).	Allowed. Normal updating, stay in CN_B.
C	Allowed. If in CN_A, stay in CN_A. If in CN_B, change to CN_A is necessary.	Allowed. If in CN_B, stay in CN_B. If in CN_A, change to CN_B is necessary.	Allowed. Stay in the registered core network.

Handover in the CS domain

Handovers between the three radio networks should follow the same principles as idle mode roaming between the radio networks. A basic difference, however, is that there is an anchor MSC and a serving MSC (MSC-B in GSM terminology) involved in the core networks. The anchor MSC will always be in the core network in which the UE is currently registered.

In order to avoid inter-MSC handovers between CN_A and CN_B, a basic requirement is that the serving MSC should be in the same core network as the anchor MSC. We should look at how this requirement can be fulfilled.

Subscribers of operator A

Subscribers of operator A should only be registered in CN_A and should only be handed over between cells belonging to radio network A or radio network C. With a suitable filtering mechanism in radio network C, the serving (target) MSC can also be chosen in CN_A. The possible handover cases for a subscriber of operator A are shown in Table A.3 below.

Table A.3: Handover cases for subscribers of operator A

to cell in →	A	B	C
From cell in ↓	A	B	C
other network	Not applicable.	Not applicable.	Not applicable.
A	Allowed. Normal handover (intra- and inter-MSC).	Not allowed.	Allowed. Intra-MSC and inter-MSC to MSC in CN_A only.
B	Not applicable.	Not applicable.	Not applicable.
C	Allowed. Normal handover (intra- and inter-MSC).	Not allowed.	Allowed. Intra-MSC and inter-MSC to MSC in CN_A only.

Roaming users

Visiting roaming users with roaming agreement with only operator A should be handled the same way as subscribers of operator A. Visiting roaming users with roaming agreement with only operator B should be handled the same way as subscribers of operator B.

Visiting roaming users with roaming agreements with both operator A and operator B are entitled to receive service via all three radio networks (A, B and C) and from both core networks (CN_A and CN_B). This would imply that all handover combinations between radio networks and core networks should be possible for these subscribers. However, from a technical and operational point of view it would simplify things if we still keep the requirement that the serving (target) MSC should be in the same core network as the anchor MSC. This limitation means that once a call has been set up in the core network that was selected in idle mode (the registered core network), the handling of that call will stay completely within that core network until the call is terminated (i.e. no inter-MSC handovers between CN_A and CN_B). It also means that if the call is started in radio network A, then it can only be handed over to cells in radio network A or radio network C and if the call is started in radio network B, it can only be handed over to cells in radio network B or radio network C. The possible handover cases are summarized in Table A.4 below.

Table A.4: Handover cases for roaming users with roaming agreements with both operator A and operator B

to cell in →	A	B	C
From cell in ↓	A	B	C
other network	Not applicable.	Not applicable.	Not applicable.
A	Allowed. Normal handover (intra- and inter-MSC).	Not allowed.	Allowed. Inter-MSC handover only to CN_A.
B	Not allowed.	Allowed. Normal handover (intra- and inter-MSC).	Allowed. Inter-MSC handover only to CN_B.
C	Allowed only if anchor MSC in CN_A. Then inter-MSC handover only to CN_A.	Allowed only if anchor MSC in CN_B. Then inter-MSC handover only to CN_B.	Allowed. a) If anchor MSC in CN_A, then inter-MSC handover only to CN_A. b) If anchor MSC in CN_B, then inter-MSC handover only to CN_B.

Handover in the PS domain

The same principles and requirements as for circuit switched handover should apply.

Cell reselection in the PS domain

The same principles and requirements as for Idle mode roaming should apply.