TECHNICAL APPROACH
Objectives | Technical Approach | Expected Achievements/Impact

The project will be carried out in the following main stages:

  1. Determination of interest and relevant target scenarios. This includes to consider:
    1. Communications environment, i.e. macrocell, microcell, indoor, etc., and user mobility
    2. Technologies deployed (GSM, GPRS, EDGE, UMTS, WLAN, WIMAX), their corresponding capabilities and functionalities, as well as their corresponding network architectures and entities.
    3. Service mix and service load (conversational, interactive, streaming, etc.).
    These activities will be carried out in WP2
  2. Development of advanced resource and QoS management algorithms, with evaluation through simulation. Focus will be placed on finding commonalities among the different scenarios considered, rather than trying to optimise algorithms and algorithmic parameters for a specific scenario. Thus, the goals of AROMA extends the mere analysis of different scenarios and will target the definition of generic end-to-end resource management criteria, facilitating their applicability in scenarios differing from those studied in detail within the project.

    These activities will be carried out in WP3

  3. Techno-economic aspects: economical analyses and evaluation of the technical outputs of the project.

    Mobile communications will continue to be one of the most dynamic and profitable market sectors in current and future economics, although it also is one of the most high demanding economic sectors from the point of view of the required investments. In such a competitive and standard-centric industrial enviroment, the economical exploitation of the solutions directed towards the optimization of the network performances are of key importance. For this reason, it is considered fundamental for the AROMA project to have the opportunity to also carry out techno-economic analyses and evaluations of the technical issues addressed by the project, investigating also the businness impacts of these solutions.

    The activities will be focused on the:

    • individuation of the value proposition and value drivers of the project,
    • definition of the possible business models,
    • economical analyses and evaluations of the business impacts in qualitative and quantitative way

    The economical evaluation may require the development of business cases based on different market scenarios and sensitivity analysis.

    These activities will be carried out in WP2

  4. Validation and demonstration of the proposed algorithms for the defined scenarios by means of a real time testbed supporting IP-based mobile multimedia applications with end-to-end QoS capabilities. These activities will be carried out in WP4

These main research topics in AROMA will be addressed within a proposed end-to-end QoS management framework aligned as much as possible with the QoS architecture envisaged in 3GPP Release 5 and 6 and other relevant IETF proposals. In this sense, it is assumed within the project that any end-to-end QoS architecture for converged 3G mobile – wired IP networks should be compliant with 3GPP UMTS QoS general framework (ref. 3GPP TS 23.107, TS 23.207).

Reference architecture

The concept of All-IP in wireless networks is commonly used to refer to those systems which include IP-based multimedia services, IP-based transport, and the integration of IETF protocols for such functions as wide-area mobility support (Mobile IP), signalling (SIP), and Authentication, Authorization and Accounting (RADIUS, Diameter). In a pure All- IP Heterogeneous network, where the IP transport would be present from a network gateway (e.g. Gateway GSN within UMTS) down to an access point (e.g. Node B in UTRAN), two different architectures can be considered:

  1. One which is close to the present architecture of 3GPP and which separates the RAN from the CN. In this case, this separation will remain and the IP RAN is connected to the CN through RNCs or equivalents [2] .Thus, IP mechanisms would be used inside the RAN and the CN but mobility and resource management within both parts could be addressed independently.
  2. The other architecture is more generic. NodeBs, Access Points (APs) or any other access technology attachment points are simply connected to an IP access network, which deals with resource management (including both radio and IP resource management). This architecture, where different radio access technologies entities are easily plugged into a generic access network, is similar to “IP2 “(IP-based IMT network Platform) [3]. Under the “IP2” approach, no clear separation exists between the RAN and the CN. Instead, a unified IP-based mobility and routing is envisaged from network gateway to Mobile Terminal (MT). Conventional separation of RAN and CN is realized functionally by the boundaries between territories where RRM and Call/Session management are applied.

These two architectures differ in the physical location of the RRM functionalities in the access network.

Both All-IP approaches could be applied to beyond 3G and wireless networks, and as according to several comments in UTRAN evolution workshop [4], there is not a general consensus about the reference architecture for All-IP wireless networks. For that reason, and taking into account the more focalised orientation of the STREP instruments, the AROMA project will assume an end-to-end reference architecture close to the 3GPP orientation. An illustration of the reference architecture for investigating performances and requirements of innovative radio resource management algorithms could be the one depicted in the next figure:

Figure 1.- All-IP mobile network architecture (all-IP UTRAN evolution for UMTS)

As it can be seen in Figure 1, in the chosen approach for “all-IP” access network, the IP transport is present from the gateway of the access network up to specific access points within each different RAN. Access Router (AR) functions are shown in the figure co-located to the RNC (or equivalent entity depending on access technology) and as the architecture considered should be compatible with 3GPP architecture, the hierarchy of GGSN, SGSN and RNC is kept. Within the RAN, IP would be used as a transport technology to connect access routers/RNCs with access point devices (i.e., Node-B, BTS, and 802.11 AP). The AP is seen as the layer 2 device that offers the wireless link connection to the mobile node.

In these All-IP wireless networks, IP can be deployed in two modes: the transport mode and the native mode. In transport mode, IP is merely used as a transport technology so as IP routing is done according to network components’ addresses (RNC, SGSN, GGSN) and user IP packets are encapsulated and transferred over this overlay network. On the contrary, IP native makes use of user IP addresses to route packets within the network. This means that no encapsulation is needed and consequently transport overhead is reduced. Thus, IP native may lead to significant improvements in terms of network efficiency and performance. It is believed that extended native use of IP in the terrestrial segment of a wireless operator’s domain that more readily allows for building a converged network with multiple access technologies.

Furthermore it can be noticed that in the “all-IP” architecture of UMTS R5, it is specified that DiffServ architecture should be supported in the different interfaces: Iur, Iub (TS24.434, TS25.426), Iu (TS25.414) and Gn. However the IP QoS control management of the access network is left open. Then in the following section issues such as IP QoS control management for a Beyond 3G system, or the interaction between the IP QoS and the RRM entities in order to provide an optimal end-to-end QoS are examined

End-to-end QoS Framework

The AROMA framework is based on the QoS framework developed in IST-EVEREST [5] but important extensions are envisaged to cover the introduction of IP RANs as well as other mechanisms to support QoS in the IP-based core network (e.g. MPLS). In the same way, different inter-working approaches among RANs are going to be further developed so as their impact on the QoS framework is assessed.

More specific, the reference QoS framework is actually based on the UMTS QoS architecture introduced in the 3GPP releases 5 for IMS services and extended to other services in release 6, [3]. Therefore, following the Policy-Based Networking (PBN) approach developed in the legacy IST-EVEREST project [1] ,the reference QoS architecture for this project will be that of a heterogeneous all-IP Network, composed mainly by two different segments: the radio access technologies (UTRAN, WLAN, WIMAX, etc..) and the Core network part composed by a pure IP network, where the IP transport is present from a gateway (GGSN) down to the RNC or even at the NodeB or access point. Notice that the proposed reference QoS architecture, shown in Figure 2, extends the 3GPP model so as to fulfil the heterogeneous all-IP Network QoS requirements. That is:

  • Policy-based mechanisms are deployed to cope with resource management in the radio access part as well as in the CN. In this sense, resource usage in the entire B3G network elements is expected to operate under a set of policies that guide the system behaviour. In the current 3GPP solution policy-based management is limited to IP QoS resource authorisation at the GGSN.
  • Dynamic QoS negotiation among all the potential RATs over which a connection may be served and the CN should be possible. In this sense, it is expected that the way QoS is offered over the overall B3G network domain and can be dynamically adapted to the radio access and core network conditions.
  • QoS management should encompass Common Radio Resource Management (CRRM) in coordinated RANs connected to the same CN. So, CRRM mechanism should be made available to the QoS management framework.
Figure 2.- Proposed reference architecture for a heterogeneous all-IP Wireless Network

The key aspects of this QoS management architecture are the following:

  • Two new functional entities are introduced to support the policy-based approach: the Bandwidth Broker (BB) and the Wireless QoS broker (WQB). The BB [6] is in charge of the control plane of the DiffServ domain while the WQB is the counterpart of the BB for the radio part of the access network.
  • Relationships between the master PDF (M-PDF) and the new entities WQB and BB is envisaged in terms of QoS negotiation. Then, session’s QoS-requirements for the whole B3G network domain are provisioned accordingly in the radio access and in the core network as a result of this negotiation. Furthermore, the M-PDF is in charge of QoS negotiation with external peer domains involved in the provisioning of end-to-end services. Inter-domain QoS negotiation can follow different approaches, [5], [8]. For instance in [5] network wide policies are established among peer domains through the exchange of updated Service Level Agreement (SLA) information.
  • Negotiation of QoS between the user (or a proxy in the Service Support Domain) and the MPDF is achieved by a policy-based service negotiation protocol (e.g. COPS-SLS [9]). After a session has been negotiated and authorised, resource activation can be committed through the usage of the PDP Context signalling as specified in UMTS although an enforcement solution from QoS entities can be also envisaged (e.g. resources in the edge DiffServ routers of the CN through the COPS-PR protocol)
  • The proposed architecture is valid for any degree of coupling among the heterogeneous RANs.

As stated above, the Wireless QoS Broker entity can be seen as the counterpart of the BB for the radio part of the access network, each entity considering the specific characteristics of the corresponding responsibility segment. The envisaged functions for the Wireless QoS Broker entity are briefly summarised hereafter:

  • Dynamic QoS negotiation among CN and RATs. Coordination is needed between the WQB and the BB, as the admission control and handover decision are submitted to different constraints in the radio part and the IP CN of the mobile access network. In the first case these constraints are related to the radio resource usage, and in the latter case to the network topology and the traffic distribution in the network.
  • Common Radio Resource Management. CRRM functions play a crucial role within the WQB. In particular, the WQB will hold the RAT selection decision function needed for the Vertical Handover decision as well as the initial RAT selection.
  • Configuration of RAN elements for QoS provisioning. As each RAN may have specific QoS mechanisms, the WQB might be responsible for the configuration of such mechanisms in order to achieve the expected behaviour. Thus, the WQB would configure QoS mechanisms in RATs elements according to a set of common policies. In the same way, CRRM functions can be configured from the WQB.

Figure 3 summarizes graphically these functions.

Figure 3.- Envisaged functions for the proposed Wireless QoS Manager

From the viewpoint of the core Network a pure IP access network is envisaged. The 3GPP does not specify a particular solution for the QoS management in the IP CN, leaving the choice to the operator among a plethora of solutions (IntServ with RSVP, MRSVP, NSIS, DiffServ, etc).

Among the different QoS solutions it is thought that a DiffServ solution with a control plane based on a BB entity, by its centralized way of management, is an appropriate choice to coordinate the radio access bearer services (radio access BS), the CN BS and the external BS (the BS provided by the neighbour domain) in order to provide an end-to-end service to the user. Moreover, its centralised architecture, articulated around the BB, allows easy interactions with the WQB for the QoS negotiation.

In addition to the coordination between the WQB and the BB for an optimal end-to-end QoS decision, the other issue examined is the rerouting after a handover and the QoS provisioning preparation before the handover along the new path.

In summary, the chosen architecture is a combination of the following elements which are inter-dependent: QoS routing (explicit routing) at the IP layer, tunnel-based mobility management (GTP and tunnel-based IP micro-mobility protocols like BCMP or HMIP), and a hierarchy of BBs. Moreover the MPLS paradigm can be complementary to the presented QoS architecture. MPLS would optimize the inter-working of IP and link layers. The IP layer would be in charge of signalling: QoS signalling between BB and ARs, IP mobility management signalling; and the MPLS and link layer would be in charge of the packet forwarding at layer 2 and the packet scheduling including QoS functions.


REFERENCES

  1.  IST EVEREST Project, 6th Framework Program of the European Community, http://www.everest-ist.upc.es.
  2.  J. Pérez-Romero, O. Sallent, R. Agustí, M.A. Díaz-Guerra, Radio Resource Management for UMTS, John Wiley & Sons, June 2005
  3.  H. Yumiba, K. Imai, and M. Yabusaki, “IP-based IMT Network Platform”, IEEE Pers. Commun., Oct 2001.
  4.  3G3GPP RAN Evolution Workshop, Toronto (Canada) 3rd-4th November 2004. ftp://ftp.3gpp.org/workshop/2004_11_RAN_Future_Evo/).
  5.  R. Ferrus et alt. “End-to-End QoS Architecture for Multi-Domain and Wireless Heterogeneous Access Networks: The EVEREST approach”. WWRF workshop in Beijing (China) February 2004
  6.  R. Neilson, J. Wheeler, F. Reichmeyer, and S. Hares, “A discussion of bandwidth broker requirements for Internet2 Qbone”, 1999.
  7.  W. Zhuang et al., “Policy-based QoS management architecture in an integrated UMTS and WLAN environment”, IEEE Communications Magazine, Nov. 2003.
  8.  IST-Project TEQUILA (http://www.ist-tequila.org)
  9.  J. Boyle, R. Cohen, S. Herzog, R. Rajan, and A. Sastry, “The COPS (common open policy service) protocol”, IETF RFC 2748, Jan. 2000.