Phenomena in 4G Networks

After introducing the main concepts related to mobility in 4G networks and the relevant methods and models linked to the evaluation of QoE, the potential effect that mobility across heterogeneous networks may have on the QoE of mobile users in 4G networks. There are three main sources for potential degradations in overall call quality: (1) changes in the underlying network conditions, (2) network handover, and (3) codec adaptation to network changes. These phenomenona will be briefly described next, as they are the core conditions studied in this work.

Changes in Network Conditions

A main characteristic of 4G Networks is heterogeneity. This translates into potentially severe changes in network conditions when seamlessly roaming across wireless technologies. The variations in packet loss, delays and bandwidth to which mobile applications are confronted with in 4G networks may lead to changes in the user's quality experience. Variations in network conditions are common, but the magnitude of the changes will increase in 4G networks, and new methods to deal with this will be needed in order to maintain an acceptable QoE. A first step towards building these new methods is to understand the impact of these variations on user experience. This chapter discusses this impact for an ongoing VoIP call when a user moves between independent wireless access networks.

Network Handovers

In Next Generation Mobile Networks (NGMN), the convergence of a multitude of different network technologies will provide the user with transparent and ubiquitous access to data and media services. The independence of network and service layers also introduces a new level of mobility. Equipped with appropriate terminals, users may move through geographical areas covered by different wireless network technologies while service access is preserved. In order to guarantee the provision of seamless connectivity especially for time-critical services like media streaming or Voice-over-IP (VoIP), sophisticated mobility-enabling protocols like Mobile IP are needed which ascertain a fast and robust roaming between heterogeneous wireless networks. This roaming between disparate wireless technologies (vertical handovers) while keeping ongoing sessions forces a new level of application level adaptation, causing an effect on user perception. This effect has not been thoroughly quantified as it is not present on existing communication systems.

Codec Adaptation

Current voice over IP codecs can adapt to variations in the underlying networks, but may not cope with the kind of changes caused by vertical handovers, as these could result in the need to change codec families (e.g., narrowband to wideband). In this case, for an ongoing VoIP call the network handover can occur in conjunction with the appropriate application layer adaptation in order to account for the newly encountered conditions. Most important, a codec re-negotiation might be enforced to meet sudden changes in network conditions. This should ideally be done in such a manner that neither the call nor the audio stream is interrupted during the handover. Thus, the continuity of the speech is maintained, resulting in an optimal speech quality.

Special attention has to be paid to these aspects as they produce new auditory artifacts that can invalidate speech quality models in the context of forthcoming communications systems, causing mismatches in the design of networks and mobile devices and applications.

IPv6 Networks and LTE

In EPC architecture there are two gateways that may be combined into a single network element; the SGW and the PGW (Figure 1). If these two gateways are separate network elements, there are two options for the protocol being used among them: GPRS Tunneling Protocol (GTP) and Proxy Mobile IP v6 (PMIPv6) .

 

Figure 1: All-IP LTE structure

The Release 8 specification identifies EPS bearers similar to the concept of PDP Context, which is defined in 3GPP's Release 7 specifications. An EPS bearer is a logical connection between a UE and a gateway, associated with a specific QoS class. As long as these Service Data Flows (SDFs) belong to the same QoS class, an EPS bearer can carry multiple SDFs.

An EPS bearer can carry both native IPv6 and IPv4 traffic in contrast to the PDP Context defined in 3 GPP Release 7. Therefore IPv4 and IPv6 stacks can be supported by a UE simultaneously as long as it's connected via a single EPS bearer.

Regarding the IPv4 scarce address resources, allocating both IPv6 and IPv4 addresses to a device does not solve the problem of IPv4 exhaustion. A service provider may therefore decide to assign only IPv6 addresses to certain devices, even when the device is able to support IPv4 and IPv6 simultaneously. In that case, NAT-PT or IPv6-toIPv4 http-proxy functionality may be required to connect these IPv6-only devices with legacy IPv4 end-points. Such a decision needs careful consideration and the issues identified in RFC 4966 need to be taken into account.

A User Equipment device (UE) obtains an IP address in one of the following two methods:

1. As part of the attachment procedure
2. Via a separate assignment procedure, such as DHCP or IPv6 Stateless Address Autoconfiguration.

The attachment procedure consists of the following steps:

1. The UE requests an attachment via sending a message to the MME, followed by an authentication procedure involving the HS S. The HSS sends to the MME, subscription data associated with the user as part of this procedure.
2. With a few exceptions, the MME is responsible for selecting the Serving and PDN Gateways that will be used for this UE. It sends a request for the establishment of the default bearer to the SGW, which forwards it to the PGW. This message exchanges results in the establishment of a GTP tunnel or a Mobile IP tunnel segment between SGW and PGW. As long as the user is attached, this segment remains up, even when the UE enters the idle state.
3. The MME orchestrates the establishment of the GTP tunnel segment between SGW and eNB and the (default) radio bearer between eNB and UE. The bearer between SGW and UE is torn down whenever the UE goes to idle state. If the SGW receives IP packets destined for the UE while it is in idle state, the SGW triggers the MME which starts a paging procedure.

When the IP address assignment is part of the attachment procedure, an IP address is allocated by the PGW to the UE according to step 2. This IP address is provided to the UE within the GTP control messages being used for establishing the EPS bearer according to step 3.

When a default bearer is established (disregarding the IP address assignment), DHCP or IPv6 Stateless Address Autoconfiguration (SLAAC) can be performed by the UE for obtaining an IP address. Therefore, a UE could receive an IPv4 address during the attachment procedure and an additional IPv6 address through SLAAC procedure.