In Part I we looked at the origins of Wi-Fi and efforts to integrate cellular mobile technology.
Beginnings: GSM Association’s WLAN Interworking Task Force
Taking note of the standardization of Wi-Fi by IEEE and the increasing uptake of Wi-Fi technology by the market, mobile operators and vendors in GSMA undertook the step of studying the potential integration of Wi-Fi as an alternate radio access method to the core network and services of mobile operators. A task force was created, called the “WLAN Interworking Task Force” which studied the topic between 2002 and 2004. The group investigated various use cases and defined six different levels of interworking between the Wi-Fi and cellular networks.
The simplest interworking scenario was what may be termed as “loose interworking” and consisted of common authentication, access control and billing. Such a combined network would allow a cellular operator to offer Wi-Fi access services to their subscribers, authenticate them using standard cellular methods (i.e. based on SIM cards in the cell phones), verify the service level subscriptions and bill them as needed. The main technical contribution here was the recommendation to use the EAP-SIM protocol over the Wi-Fi networks to authenticate cellular subscribers with SIM-based Wi-Fi enabled handsets.
Progressively tighter levels of interworking were defined in terms of increasingly seamless and integrated access to cellular networks and services. For example, in the loose interworking scenario, the subscriber can access the cellular operator’s core network and service network, but some services may not be accessible over the Wi-Fi radio. This limitation is removed in the second level of interworking. The second level of interworking allows “Service Continuity”, meaning that all the services that a subscriber has subscribed to must be accessible by the subscriber via the Wi-Fi network. This would include voice-mail, texting, mobile TV, and any other services offered by the cellular operator.
A limitation of the second level of interworking is that while all services may be accessed via the Wi-Fi network, an ongoing session that a user has established over the cellular radio may not, due to user mobility, seamlessly handover to a Wi-Fi network. In other words, the ongoing cellular radio session would be torn down and a new Wi-Fi session would have to be established. This would seriously impact the quality of the user experience. This limitation is removed in the third level of tighter interworking, which calls for so-called “session continuity”. This would make the experience of a user moving between Wi-Fi Networks and Cellular Networks as seamless as if he or she were still connected to the cellular network.
Shown below is a figure from GSMA PRD (Permanent Reference Document) SE.27, illustrating various integration scenarios, as understood at the time.
GSMA’s WLAN Task Force completed the study in a comprehensive manner and identified the need for standardization within 3GPP. This resulted in the development of a number of standards, based essentially on the loose integration of Wi-Fi networks with the 3GPP core network and can be viewed as an evolution of integration scenario #3 in the above figure. The detailed integration solutions depend upon whether the core network is a UMTS Core network or Enhanced Packet Core (EPC) network. The former set of standards is referred to as IWLAN (Integrated/Interworked WLAN) standards and the latter as EPC standards.
Integrated Wireless LAN (IWLAN) Standards
The IWLAN standardization work commenced with an initial feasibility study, which included some of the findings of the GSMA’s WLAN Task Force. It resulted in a 3GPP Technical Report, TR 23.234, latest version of which is v10.0.0, produced in 2011. The report essentially identifies a number of interworking scenarios, numbered 1 through 6 as follows:
Scenario 1 – Common Billing and Customer Care
Scenario 2 – 3GPP system based Access Control and Charging
Scenario 3: Access to 3GPP system PS based services
Scenario 4: Service Continuity
Scenario 5: Seamless services
Scenario 6: Access to 3GPP CS Services
The last scenario has not been pursued in standardization, but the others led to a number of 3GPP specifications, listed below:
TS 23.234: Scenarios 1 and 2: Common Billing, Access Control and Charging and Access to PS Services.
TS 23.327: Scenarios 4 and 5: Service Continuity and Seamless Services – Single Radio Case and Mobility
TS 23.261: Scenarios 4 and 5: Service Continuity and Seamless Services – Dual Radio Case and Flow Mobility
TS 23.234 provides a solution for Interworking Scenarios 1 and 2. The figure below is a simplified version of a figure from 23.234, wherein the 3GPP AAA server performs the necessary user authentication and authorization for both WLAN access and 3GPP services. Two different types of IP-Access Services are provided to the user, namely “3GPP IP access” and “Direct IP access”. The former refers to access to 3GPP packet data services, such as MMS, mobile video, etc. as well as internet services, whereas the latter refers to direct access to internet or intranets.
One of the main goals of the IWLAN solutions was to achieve authentication without manual user intervention, such as entering a username-password, as is common is many Wi-Fi networks. This is made possible by developing authentication protocols based on the use of SIM cards, which are already provisioned in the 3GPP handsets. In addition to providing authentication in a manner transparent to the user, SIM based authentication methods are also familiar to 3GPP operators and provide the same level of security as 3GPP devices.
Since the SIM based authentication is now done via WLAN networks, which are essentially IP Networks, the basic 3GPP authentication protocols are modified and are known as EAP-SIM, EAP-AKA and EAP-AKA’ protocols, which were standardized by the IETF. IP Networks also use certificate-based authentication methods, which are also standardized as EAP-TLS and EAP-TTLS protocols for WLAN based authentication.
Regarding IP access services, the 3GPP-IP-access is provided by two functional entities called WAG and PDG. As the name indicates, the PDG is a gateway to a specific Packet Data Network, such as the Internet or an operator service network. Clearly, the 3GPP network may support multiple PDNs and hence multiple PDGs, for different types of services. The WAG implements a typical gateway function on the user side, by connecting the user to one of the possibly several PDGs. It also acts as a firewall and implements operator policies, which are downloaded from the 3GPP AAA servers.
Finally, WAG performs charging related functions, as set by the operator, and communicates with the 3GPP charging systems, of which there are two types, namely offline (for post-paid customers) and online (for pre-paid customers and for checking spending limits).
While the solutions presented in TS 23.234 allow WLAN UEs to access 3GPP core networks and services, the radio connection cannot be dynamically switched between Wi-Fi and 3GPP access networks. These solutions are standardized in TS 23.327, which describes mobility between Interworking WLAN Networks and 3GPP networks.
The mobility function is essentially based on an IP-level mobility management protocol called DSMIPv6, which is standardized by IETF. This protocol is implemented in an entity called HA (Home Agent) in the core network of the home 3GPP network and in a peer entity called DSMIPv6 client in the UE. The UE has a single IP address (for the purposes of mobility management), which is called Home Address (HoA) and a Care-of-Address (CoA) which changes as the UE attachment is changed between IWLAN and 3GPP radio interfaces. Changes in CoA address are synchronized between the UE and the HA by exchanging the so-called “Binding Updates” messages. These are sent over the logical interface H1 shown in the above figure. It is supported over the chain of physical interfaces Uu/Um, Iu_ps/Gb, Gn and H3 when connected over the 3GPP access network and over Ww, Wn, Wp and H3 when connected over the WLAN Access Network.
DSMIPv6 mobility protocols allow handover from 3GPP access to WLAN access or vice versa. However, in the current version of the standards, only the UE can initiate such a handover procedure. This is based partly on the rationale that the UE has a better knowledge of the WLAN radio networks. However, there are some initiatives in the 3GPP standards organization currently that are seeking to standardize network-initiated handovers as well, since the network has a more comprehensive knowledge of the network congestion state. Although the HA function is shown as a separate function in the above figure, it is often collocated with the GGSN.
Enhanced Packet Core (EPC) Standards
The above solutions have two basic limitations. The first limitation is that the HA is not connected to policy and QoS management entities in the core network, such as PCRF. This prevents advanced policy and QoS based management of the IWLAN-3GPP mobility. These limitations are removed in the case of integration of WLAN into EPC core networks, which is described in the next section.
The second limitation is that the above solutions restrict the UE to have only a single radio connection at any given time, namely either to the WLAN or 3GPP radio interface. Modern smart phones allow simultaneous connectivity to both radio interfaces, which raises the possibility of managing 3GPP and WLAN interworking at an individual IP-Flow level. That is, it should be possible to support certain IP-Flows on the 3GPP radio interface and certain others on the WLAN radio interface, based on criteria such as QoS requirements, user subscription, type of user equipment, etc. Furthermore, it could also enable dynamic switching of individual IP-Flows from one radio interface to another. The figures below illustrate the situation.
The EPC standards for 3GPP and Non-3GPP interworking introduce a new class of non-3GPP access networks, namely Trusted Non-3GPP Networks, with the word “trust” referring to trust by the operator (and not necessarily by the user). Accordingly, Trusted Wi-Fi Networks imply that the Trusted Wi-Fi access points are deployed and managed by the Operator, so that UE can connect to the Wi-Fi Network directly using the radio interface without requiring any additional security measures.
In contrast, un-trusted Wi-Fi Networks do not have any trust relationship to the operators, so that the operators require that the UE establish a secure tunnel (i.e. IPSec tunnel) to a trusted node in the operator core network. Typically, such a node is a PDG in UMTS core networks (as in IWLAN architectures) and ePDG in EPC core networks. Shown below are two simplified architectures of an EPC core network with 3GPP as well as Trusted and Un-Trusted non-3GPP Access. Other architectures are also possible and are documented comprehensively in TS 23.401 and TS 23.402.
Note that here the PGW includes a HA functionality and that PCRF is connected to various gateway functions, each of which has a PCRF or its functional equivalent to enforce the operator policies.
Shown also is the ANDSF functionality, which is a critical one for Cellular-Wi-Fi interworking from an operator policy point of view. Currently, most smartphones choose and camp on to Wi-Fi networks based on explicit user preferences or preconfigured preferences, already stored in the UE. It was clear that if operators were to offer Wi-Fi access as an integral part of the access offerings, they needed to be able to install operator policies on the UE and also be able to change them dynamically, as the conditions may change. To achieve this, the framework of ANDSF was standardized. It essentially consists of an ANDSF server in the operator network, which stores the operator policies regarding discovery and selection of Wi-Fi access. For example, it contains discovery information of Wi-Fi hotspots based on the location of a UE. Regarding selection of Wi-Fi Hotspots, the policies may specify that certain Wi-Fi hotspots are preferred at certain locations and/or certain times of day, or for certain types of applications, such as mobile video etc. These operator policies can be transferred to the UE via the S14 interface using communication procedures based on device management procedures, originally developed by the OMA organization.
Interworking between 3GPP and non-3GPP networks essentially consists of mobility of IP-Flows between the 3GPP and non-3GPP networks. A number of cases of such mobility can be distinguished depending on the following aspects: (1) mobility is on a per IP-Flow basis or per all IP-Flows associated with a PDN connection; (2) mobility is Seamless or Non-Seamless, with Seamlessness defined as preservation of the IP-address of the UE during the mobility process. Different combinations of these two fundamental aspects result in a number of scenarios, such as Wi-Fi offload, referring to mobility of IP-Flow(s) from 3GPP to Wi-Fi networks, and handovers, referring to mobility of all IP-Flows associated with a PDN connection etc.
The 3GPP standard TS 23.401 describes Seamless and Non-Seamless Handover solutions between 3GPP and Non-3GPP access networks, wherein GTP is used as the protocol for the Handover over the interfaces S2a, S2b and S5. Similarly, TS 23.402 documents similar solutions for the cases where PMIP and DSMIP are used for mobility.
Finally, TS 23.261 describes the solutions for Seamless IP-Flow Mobility using DSMIP protocols. As mentioned below, this allows for selective assignment of different IP-Flows to different access networks and includes Seamless Wi-Fi Offload as a special case.
In all cases listed above, the mobility is triggered by the UE and not by the network. Efforts are also being made to standardize network-triggered mobility procedures, since the network is often more knowledgeable about the overall network usage and congestion state than the UE.
This concludes a review of past efforts towards cellular/Wi-Fi integration. However, with the advent of new device types and applications, the desire for increased integration is more salient now than ever before. In Part III of this series, we’ll be looking at current and future efforts, and how they are setting the stage for the heterogeneous networks of the future.
Source: https://www.edn.com/design/communications-networking/4390906/Cellular-Wi-Fi-Integration-A-comprehensive-analysis-Part-II?page=3 – Prabhakar Chitrapu, Alex Reznik, Juan Carlos Zuniga- 07.23.2012 July 23, 2012