VoIP in 3G Networks: An End-to- End Quality of Service Analysis
Renaud Cuny
1, Ari Lakaniemi
21
Nokia Networks
P.O.Box 301, 00045 Nokia Group, Finland renaud.cuny@nokia.com
2
Nokia Research Center
P.O.Box 407, 00045 Nokia Group, Finland ari.lakaniemi@nokia.com
Abstract-- This paper presents the results of a Quality of Service (QoS) study for VoIP service over 3G WCDMA networks. An end-to-end simulation platform has been used for this purpose. The simulations have been run using Adaptive Multi-Rate (AMR) speech codec at 12.2 kbit/s with combination of RTP, UDP and IPv6 protocols. The simulated transmission path includes two radio links (uplink and downlink), connected with a packet switched core network and UTRAN Radio Access Networks with several different radio transmission conditions.
Furthermore, RObust Header Compression is applied in both radio links. The results include buffering statistics, end-to-end delay estimates, and packet loss statistics.
I. I NTRODUCTION
During the last few years, the voice over data network services have gained increased popularity. Quick growth of the Internet Protocol (IP) based networks, especially the Internet, has directed a lot of interest towards Voice over IP (VoIP).
The VoIP technology has been used in some cases, to replace traditional long-distance telephone technology, for reduced costs for the end-user. Naturally to make VoIP infrastructure and services commercially viable, the Quality of Service (QoS) needs to be at least close to the one provided by the Public Switched Telephone Network (PSTN). On the other hand, VoIP associated technology will bring to the end user value added services that are currently not available in PSTN.
On the other front, the current development in the cellular radio network technologies are paving the way towards IP capable radio networks. The so called Third Generation (3G) cellular networks, developed and standardized by the Third Generation Partnership Project (3GPP), will provide IP over wireless services, enabling therefore also VoIP. In current cellular systems, e.g. in GSM, the telephony service is based on circuit switched approach. This service is currently highly optimized for transmission of voice, thereby providing good speech quality and good spectral efficiency. However, carrying VoIP will be also possible in 3G WCDMA networks, e.g.
3GPP release 5, and may be of special interest for the mobile network operators for multiple reasons: Firstly, as the bandwidth for individual flows in packet switched domain is not reserved in advance, the multiplexing effects should bring significant capacity savings. Secondly, VoIP service will be supported by the Session Initiation Protocol (SIP), which is a text-based protocol, similar to HTTP and SMTP, for initiating
interactive communication sessions between users [1]. Such sessions can include voice, but also e.g. video, chat, interactive games, and virtual reality. Finally, the convergence towards packet switched and IP technology may convince mobile operators to go for solutions that are truly all-IP in order to simplify network interconnection and network management.
Naturally, for wide end user acceptance and deployment, the VoIP service is required to provide similar perceived voice quality as provided by current highly optimized GSM networks. The challenges for achieving this include typical VoIP related QoS problems, such as packet loss, delay, and delay variation (i.e. jitter), as well as additional overhead brought by the VoIP protocol stack. Therefore the end-to-end VoIP QoS should be studied and evaluated carefully. As an example, it is likely that the packet switched technology, although managed by e.g. Differentiated Services [2], will generate more delay and jitter than the circuit switched technology. Further additional delay and jitter may be caused by the packet segmentation in the radio interface. The end-to- end delay is likely to be close to the maximum delay still providing acceptable conversational quality (around 250- 300ms [3]), extra attention needs to be paid to jitter: too much jitter for a voice stream may be problematic since basic jitter compensation methods may not apply very well or have limited effects. So one important issue to investigate is whether the jitter in 3G networks, will have negative impact on the end-user perceived voice quality.
This paper is organized as follows. Section II presents in details the end-to-end VoIP simulator used for this study. Each component of the tool is described in detail. Section III presents the simulation results focusing on packet loss ratio and end-to-end delays. Finally, the conclusion in section IV summarizes the main finding of this study and points out the areas that could be investigated further.
II. E ND - TO - END V O IP SIMULATION
Protocols used by the VoIP over 3G can be roughly divided
into two categories: signaling related protocols and media
related protocols. Although the signalling protocols, such as
SIP, are very important part of a VoIP system, in this study we
concentrate only on media related protocols and transmission
of media data.
To run the end-to-end simulations we developed a VoIP speech simulator application for modeling the telephony application and protocol layers from application down to IP and PDCP. The lower layers required for radio link and core network modelling were simulated using external simulation tools and the resulting network conditions were applied in the VoIP speech simulator using error pattern files. The different components of the simulation chain are described in detail in the following subsections.
A. Speech application
On application level we assumed usage of Adaptive Multi- Rate (AMR) speech codec, which is a mandatory codec for conversational speech services within 3G systems. For all simulation runs we selected usage of AMR 12.2 kbit/s mode with DTX functionality enabled, and employed bandwidth efficient mode of the AMR RTP payload format. This implies that during talk spurts the source generates 32-byte speech payload at 20 ms intervals, while due to DTX during silence periods we will have 7-byte payload carrying Silence Descriptor (SID) frame at 160 ms intervals.
We further assumed the typical VoIP protocol stack employing Real-Time Transport Protocol (RTP) encapsulated in User Datagram Protocol (UDP), which is further carried by the IP. The combination of these protocols introduces total of 40 bytes header data when using IP version 4 (IPv4), and bytes header when using IP version 6 (IPv6). We selected IPv6, which has two implications: the size of an IP packet carrying one AMR frame will be either 92 bytes (speech) or 67 bytes (SID), and we need to enable UDP checksum because the IPv6 header does not include a checksum of its own but the most critical fields of the header are covered as part of the UDP pseudo header.
Protocol layers below IP follow the 3GPP release 5 specifications, as illustrated in Figure 1.
Figure 1 3GPP Protocol stack
B. Robust Header Compression (ROHC)
When operating in the bandwidth limited 3G networks it is important to use the radio band as effectively as possible, and header overhead up to 60 bytes can seriously degrade the spectral efficiency of a VoIP service over such link. The
RObust Header Compression (ROHC) protocol [4] has been developed to tackle this problem. ROHC provides link-based compression of IP/UDP/RTP headers, in best case down to 1 byte. The effective compression makes use of the fact that majority of the fields in the combined IP/UDP/RTP header either remain constant or introduce constant change throughout a session. However, the maximum compression mentioned above can only be reached when imposing some limitations, a more typical compressed header size would be three or four bytes. The ROHC operation is based on synchronized compression (at the sender site) and decompression (at the receiver site) contexts. The decompression context is initialised by transmitting full IP/UDP/RTP headers in the beginning of the session. Also irregularities in the transmitted stream e.g. by DTX operation or lost packet can introduce compressed headers slightly larger than in the optimal state. In error prone transmission conditions a feedback mechanism is important part of robust compression operation, enabling recovery in case the synchronization between compressor and de-compressor is lost. The ROHC protocol was implemented in our simulator.
The ROCH in R-MODE is assumed on both radio links, providing feedback mechanism to enable safe convergence to optimal compression state. We also assume that ROHC Context Identifier is transmitted as a part of the compressed packet. These settings imply that the minimum size of a compressed IP/UDP/RTP header is four bytes.
C. Radio network modeling
The model for the radio network included the actual radio link, processing in layers below PDCP and access transport in UTRAN. The radio link error patterns were prepared using a separate WCDMA system simulator. Three different radio conditions were investigated, introducing frame error rates (FER) of 1%, 3% and 5%. Additionally we also included error-free case in the set of simulation conditions. Different error patterns were prepared for both uplink (UL) and downlink (DL), and the error patterns were obtained from a traced terminal that was moving along a predefined route.
For UL radio network we assumed processing and transport delay of 36 ms, and for DL radio network the corresponding delay is 49 ms. Note that 36+49=85 ms is the lower limit for the time before the ROHC compressor can receive a feedback message from the decompressor regarding a specific packet.
This delay is significant in such a way that in beginning of a stream the ROHC decompressor context needs to be initialized by sending full headers, which will be sent until a feedback message indicating successful decompressor context initialization is received. A similar situation can occur also if the decompression context gets corrupted for some reason, e.g.
hard handover or excessive amount of transmission errors.
However, for this work we assumed that no ROHC decompressor context re-initialization is required during a session.
L1 RLC PDCP
MAC E.g., IP,
PPP Application
L1 RLC PDCP
MAC L1 UDP/IP GTP-U
L2 Relay
L1 UDP/IP
L2 GTP-U E.g., IP, PPP
3G-SGSN UTRAN
MS
Iu-PS
Uu Gn Gi
3G-GGSN L1
UDP/IP GTP-U
L2 L1 UDP/IP
GTP-U
L2 Relay