Key Technologies of 3G WCDMA Network Sharing

WCDMA 3G network sharing technology allows different telecommunication network operators to share part of the elements of the WCDMA network. These operators can, according to their own strengths, cooperate on construction and/or operation of the WCDMA networks by network sharing. According to the results of relevant market investigation, operators may be able to save about 20% to 40% of costs at the stage of network coverage and expansion by using network sharing[1]. Thus this technology gains worldwide attention from the main operators and equipment vendors.

    The initial 3GPP specifications never took into consideration the possibility of sharing part or the whole network by two or more independent operators. In Release 99 and Release 4, User Equipment (UE) could handover to other optional Public Land Mobile Network (PLMN), which can be regarded as a limited support of network sharing. Release 5 introduced the concept of  Iu-flex and the feature of sharing UMTS Terrestrial Radio Access Network (UTRAN) under the connected mode. Release 6 proposed 4 network sharing solutions, and had an emphasis study on enhancement of the UTRAN sharing support function.

    On the 26th meeting of 3GPP Technical Specification Group of Radio Access Network (RAN), held in Athens in December 2004, the work item on network sharing was proposed to close[2], which indicated, the network sharing technology basically became mature.

1 Four Network Sharing Solutions
There are 4 network sharing solutions: site sharing, common network sharing (RAN and part of Core Network (CN) sharing), geographically split network sharing and RAN sharing (i.e., UTRAN sharing)[1,3]. The selection of the solution depends on specific market circumstances, traffic anticipation and allowance status. One or multiple operators or an independent third party may own the shared network elements.

    Site sharing usually includes the sharing of the site location, the antenna tower, antennae, power equipment and transmission equipment. One site may be shared by two GSM operators, or by two WCDMA operators, as well as by one GSM operator and one WCDMA operator. The site sharing saves expenditure on site acquisition, building projects, yearly site rent and site operation. The site sharing is suitable for both, densely populated and rural areas.
Common network sharing refers to the cooperation of multiple operators to construct and operate a common 3G network, including the RAN and part of CN. This solution helps to improve the spectrum efficiency. Each operator who’s sharing the network can deploy one frequency to provide initial coverage. It is very cost effective to use this solution in a large area with limited traffic expectations. The solution also allows network sharing between an operator without the 3G license and an operator with the 3G license.

    In the solution of geographically split network sharing, each operator owns its independent 3G network. The operators have different geographical coverage, and expand the coverage of their 3G services by allowing national roaming. The traffic in the shared area is pooled on individual operators’ frequencies, which improves the network efficiency. In rural areas, This is a solution of a long-term coverage expansion. The operators still individually maintain their networks, but those with larger coverage area may provide extra coverage for the operators with limited coverage. This requires a signed agreement of either mutual or one-sided roaming.

    As for RAN sharing, operators share the RAN, including the Radio Network Controller (RNC) and the base stations. Figure 1 illustrates that one physical RNC is divided into two or more logical RNC, belonging respectively to different operators. Each logical RNC has its own PLMN code, carriers and frequencies. The RNC is shared physically but logically separated. The RAN sharing can be introduced in the RNC software, which can expand the frequency and the radio network capacity. This solution is typically applied in densely populated areas to meet the coverage requirements of 3G networks. Besides, it can also be used to provide joint coverage of areas outside licence requirements. Each operator can connect the shared RAN for individual coverage.

2 Key Technologies of Network Sharing

2.1 Iu-Flex Function in Release 5
Before Release 5, one RAN node is only permitted to connect to one CN. This has some drawbacks. The hardware resources are apparently wasted when a RNC, compared with the Mobile Switching Center (MSC)/Serving GPRS Support Nodes (SGSN), has a large capacity. Besides, considering the signaling traffic in network, if there are more MSC/SFGSNs in one network, more registration updates will appear, and the related inter-MSC/SGSNs updates will cause extra signaling overhead in the CN and at the radio interfaces.

    Release 5 introduces the function of Iu-Flex. So the RNC is able to support intra domain connection of  one RAN node to multiple CN nodes, allowing the RAN node to route information of the corresponding Circuit Switching (CS) domain or Packet Switching (PS) domain to different CN nodes. In this way, multiple MSCs or SGSNs share the loads, and the hardware utility is further improved.

    There is one or more MSC or SGSN service areas concentrated in a pool area (see Figure 2) and UE may roam in the area without changing the service node of the CN. One pool area can serve several parallel CN nodes, and it can be independently configured to CS and PS domain. From the view of RAN, the pool area includes all Location Areas/Route Areas (LAs/Ras) of one or multiple RNCs served by a group of parallel CN nodes. It is also possible for this group of parallel CN nodes to serve any of the LAs/Ras out of the pool area or other pool areas. Besides, this group of CN nodes may refer to corresponding MSC or SGSN pools. Multiple CN nodes in one pool area may realize load balance, and the service range they provide is wider than that provided by a single CN node. This result in the reduced  inter-CN node updates, handovers and relocations, and it reduces the Home Location Register (HLR) update traffic.

    The following is the basic working procedure of the Iu-Flex:

(1) Setting up the SGSN
    When UE executes GPRS attachment, RNC will select an appropriate SGSN. The SGSN encodes its identifier, the Network Resource Identifier (NRI), to Packet Temporary Mobile Subscriber Identifier (P-TMSI) to inform UE, RNC and SGSN of the mapping relationship of the IMSI and NRI. RNC and SGSN may correspondingly route packets.

(2) Record of SGSN
    When UE is under the Packet Mobile Management (PMM)-connected mode, RNC will keep the IMSI-NRI mapping. But when UE is in the PMM-idle mode, RNC will delete the data of UE (no routes needed for packets either from or to UE anymore). When UE is PMM-connected again, the NRI of its SGSN will be sent to the RNC again.

    It should be noted that Iu-Flex doesn’t permit different CN operators to share the same RAN yet, however, the enhanced function of network sharing support of Release 6 has already implemented it.

2.2 UTRAN Sharing Mechanism Under Connected Mode in Release 5
In Release 99 and Release 4, the CN and the UTRAN all provide the access limitation mechanism to current PLMN and other PLMNs when the UE is in idle mode. This mechanism can be used for network sharing. Based on the roaming agreements, it can set different access restrictions on different UEs, However, insufficient mechanisms are specified to provide similar access restrictions in Connected Mode. The Shared Network Area (SNA) introduced into Release 5 solves the problem of UTRAN sharing in connected mode[4]. SNA-Id is the identifier of SNA and is used to identify the area that includes one or more Las, i.e. SNA. SNA endows UE in connected mode with a right to access the shared network. The SNA-Id is composed of PLMN Identifier and Shared Network Area Code (SNAC). SNA-Id=PLMN-Id||SNAC, and SNAC is defined by the operators.

    The shared network access control function supports the CN to request UTRAN to assign an access control  to the LAs of UTRAN and the neighboring networks through UE. The function is based on the entire PLMN or SNA. One SNA corresponds to one or more LAs. One or more LAs belong to one independent PLMN, and the UE access can be controlled in this PLMN. In order to implement shared network access control of UTRAN or its neighboring systems, UTRAN needs to know whether the related LAs belong to one or more SNAs. If it is necessary to do access control on a certain UE, CN will provide UE with the access information of SNA. The information designates the UE which PLMNs and/or SNAs are allowed to access. The UTRAN determines whether to allow the UE to access the LAs or not on the basis of whether the LAs belong to the PLMN or the SNA. If it doesn’t permit the access, the UTRAN will request the CN to release the relevant resources. Due to the access control of the shared network, both the release request and the relocation request are initiated through the Iu interface.

2.3 Network Sharing Support Enhancement Mechanism in Release 6
The UTRAN sharing mechanism in Rlease 6 allows different operators to share the network elements of RAN, as well as the radio resources.

    RAN sharing has two models [5, 6]:
    (1) Besides RAN sharing, MSC and SGSN are also shared. This configuration is called the configuration of Gateway Core Network (GWCN), as illustrated in Figure 3.

    (2) Only RAN is shared. This configuration is called the configuration of Multiple Operators Core Network (MOCN), as illustrated in Figure 4.

    Network sharing is an agreement between the operators. For the users, it is transparent. Therefore, the supporting UE has to discriminate between the CN operators available in a shared RAN and that these operators can be handled in the same way as operators in non-shared networks. System broadcast information in the shared network should include the information of each cell and each CN operator in the shared RAN. The supporting UE decodes the system broadcast information to determine available CN operators in the shared network. Non-supporting UEs ignore the related information about network sharing in the system broadcast information.

    When a UE performs initial accesse to a shared network, it will select one of the available CN operators to serve it. For MOCN, RAN routes the UE’s initial access to the shared-RAN to one of the available CN nodes. When the initial access to the shared network is completed, UE will not handover to other available CN operators as long as the chosen CN operator can serve it.

    In conventional networks, it is always the same CN operator to provide both CS and PS services to the UE. In a shared network, supporting UEs shall behave as UEs in conventional networks with respect to registration with CS and PS domains. The UE uses the information stored in Subscriber Identification Module (SIM)/Universal Subscriber Identification Module (USIM) to attach the CN from which UE was detached. For the supporting UE in shared network, the information stored in SIM/USIM indicates the CN operator from which it was detached. However, for the non-supporting UE in shared network, the information stored in SIM/USIM indicates Common PLMN.

    The supporting UE shows the PLMN-ID it has been registered. In the shared network, it is the PLMN-ID of the selected CN operator. Using shared Visitor Location Register (VLR)/SGSN should not result in  service restriction, e.g., roaming restriction. According to the VLR/SGSN number, HLR can judge whether the user is roaming in the H-PLMN or V-PLMN, and for GWCN, the shared VLR/SGSN will be assigned a specific number from each supported H-PLMN. For example, the shared VLR/SGSN has multiple numbers, and the VLR/SGSN number assigned by the serving CN operator is used when it exchanges signaling with the HLR.

    The supporting UE not only selects the CN operator, but also provides its PLMN-Id to the network for routing. The information of the available CN operator is transmitted by the RNC via the system broadcast information. If the system broadcast information is transmitted to the supporting UE in the dedicated signaling, the RNC is able to indicate the PLMN-Id of the connected CN operator (if the PLMN-Id is included in the signaling). As for the
non-supporting UE, the RNC is able to indicate the PLMN-Id of the common PLMN (if the PLMN-Id is included in the signaling).

    When transmitting the initiam layer 3 signaling, the RNC,  for the supporting UE, indicates the selected CN operator to CN. The selected operator is indicated in the Radio Resource Control (RRC) signaling or implied in the connected signaling. As to the non-supporting UE, RNC cannot show the selected CN operator to CN.

    When the UE accesses MSC at the first time, i.e., when there is no VLR entry for this UE, the MSC verifies whether the UE belongs to one of the operators sharing the MSC or their roaming partners. For MOCN, MSC possibly fails to provide service for the UE, and at this time, the UE has to relocate to another MSC belonging to another CN operator. At last, the MSC/VLR that offers service for the UE will assign one NRI for the UE to permit the RAN to route any subsequent UE accesses to the serving MSC/VLR.

    When the UE accesses SGSN at the first time, i.e., when the UE is unknown to the SGSN, the SGSN verifies whether the UE belongs to one of the operators sharing the SGSN or their roaming partners. The SGSN implements it by obtaining International Mobile Subscriber Identity (IMSI) information from another SGSN or the UE. The SGSN will select a CN operator, unless the old SGSN or the UE has already indicated one. The SGSN also stores the identifier of this CN operator. For MOCN, the SGSN possibly fails to provide service for the UE, and at this time, the UE has to be relocated to another SGSN belonging to another CN operator. At last, the SGSN that serves the UE will assign one NRI for the UE to allow the RAN to route any subsequent UE accesses to the serving SGSN.

3 Conclusions
Network sharing is getting wide attention from operators and equipment vendors. Based on the latest findings of 3GPP, this paper analyzes and summarizes its implementation solutions and key technologies. The study on network sharing support enhancement of UTRAN in Release 6 network has basically finished in December 2004. The study results have approved by the related 3GPP work group. It is expected to be written into corresponding protocols.

References
[1] Ericsson White Paper. Shared Networks for WCDMA[Z]. 2003, 8.
[2] 3GPP RP-040428 Status Report for WI to TSG[S]. 2004, 12.
[3] 3GPP TR 22.951 V6.1.0 Service Aspects and Requirements for Network Sharing[S]. 2003, 3.
[4] 3GPP TS 23.236 v5.2.0 Intra-domain Connection of RAN Nodes to Multiple CN Nodes[S]. 2002, 3.
[5] 3GPP TR 23.851 V6.1.0 Network Sharing Architecture and Functional Description[S]. 2004, 6.
[6] 3GPP TS 23.251 V6.1.0 Network Sharing Architecture and Functional Description[S]. 2004, 9.

Manuscript received: 2004-12-16