IEC Newsletter
November 2006, Volume 1 back to index
Enterprise IP Phones: The Voice of the Future
By Dr. Carsten Sydow and Daniel Hartnett, Infineon Technologies AG

What is the benefit of an Internet protocol (IP) phone in the enterprise network? A phone connected to an IP network—instead of the conventional public switched telephone network (PSTN) line—that also delivers data to the user sounds like a winning proposition. And it is. IP phones offer a host of new applications and enabled services that turn the once voice-only console into a powerful networking device. Audio- and videoconferencing, personalization, and interfaces to wireless devices such as Wi-Fi and Bluetooth are only examples of the enhancements to end users' experience. Gigabit Ethernet support and enhanced security mechanisms can put even the strictest corporate IT manager at ease. And to top it all, when designed right, IP phones provide far better voice quality than analog phones.

IP phones are connected via Ethernet to an IP network. IP networks transport the data in packets, offering high bandwidth for all participants in the network. Figure 1 shows a typical scenario. Traditional telephone networks, on the other hand, transport voice on a dedicated logical circuit and offer a fixed but small bandwidth for each participant.

Figure 1 IP Phone System

Gigabit Ethernet
Only a few years ago, it was hard to imagine users requiring more than 100 Mbps (fast Ethernet) speeds on the desktop. As networks scale upward, however, the requirement to support gigabit Ethernet (GE)—or 1,000 Mbps—to the desktop is fast becoming reality. With applications such as videoconferencing and streaming, it is only a matter of time before this bandwidth will be utilized. IP phones are not the endpoint of this architecture and, as such, should not represent a bottleneck for the user. Next-generation IP phone system on chips (SoCs) already embrace this trend by integrating a GE switch, which is capable of not only of switching GE traffic, but also of implementing many of the security features mentioned in the security section to follow.

Wideband Telephony
The human ear is capable of hearing frequencies from 20 Hz to 20 kHz. Traditional telephone systems transmit an acoustic bandwidth of 300 Hz to 3.4 kHz. This "narrowband" bandwidth is a compromise between speech intelligibility and bandwidth consumption. The PSTN transmits that data coded with the G.711 codec and sampled at 8 kHz.

"Wideband" telephony specifies a transmission range of 150 Hz to 6.3 kHz. While this is not CD bandwidth (20 Hz up to 20 kHz), the increased bandwidth compared to narrowband offers significantly increased intelligibility and a much more natural sound. Wideband telephony was standardized for the integrated services digital network (ISDN) with the G.722 codec but never really enjoyed wide proliferation.

Since IP phones already have powerful signal processing capabilities for narrowband speech compression algorithms, wideband codecs can easily be handled by the voice engines in IP phones. If the analog-to-digital and digital-to-analog converters support a 16 kHz sampling rate, wideband telephony on an IP phone comes with relatively low additional overhead. Other factors driving the development of wideband telephony are personal computers (PCs), mobile phones, and the new Digital Enhanced Cordless Telecommunications (DECT) standard.

With the wider deployment of IP telephony, there are now many more codecs in the voice over Internet protocol (VoIP) space. Supporting these and customer-owned codecs will be a challenge for IP phones moving ahead, as their SoCs will be required to support not only the traditional G.7xx codecs, but also a plethora of royalty-free and quality-enhancing codecs such as Internet low-bit-rate codec (ILBC), G.722.1, adaptive multi-rate (AMR), and MP3. This will be a big challenge in terms of storage and scalability for the SoC.

While it is a major problem to get the bandwidth required for transmission of a stereo or a multi-channel signal in traditional telephony, the IP transmission medium offers a sufficient bandwidth that can easily be used for multichannel applications.

Web Radio
Web radio is another feature that can be realized on an IP phone. This requires a broadband or Ethernet connection, the necessary software, a user interface, an audio decoder, and an acoustic front end. In some cases, the connection of an AC97 codec makes the sound experience even more exciting. The above features are already part of some IP phone systems and in many cases are directly supported by the central processor in the SoC.

Video Telephony
IP phones can and will also support video telephony. The challenge for video phone designers is to create a product that delivers what users perceive as high-quality video using state-of-the-art compression algorithms such as Moving Pictures Experts Group 4 (MPEG-4) (H.263 and H.264 protocols). Video phone deployment has traditionally been stalled by high system costs (including TFT display) and the processing-power cost of delivering high-quality, highly compressed video over bandwidth-limited networks. For video telephony, the video signal has to be transported in addition to the audio signal. The picture of one talker is captured by a camera, coded for bit-rate reduction, transported to the remote participant, decoded, and displayed. Voice and video must also be synchronous. The audio is already processed by an IP phone, and the IP connection offers enough bandwidth to transport video as well. The camera, the display, and the video codec have to be added.

Security Mechanisms
Unlike its PSTN predecessor, the IP phone represents an integral part of the enterprise network, not its endpoint. For this reason, the phone has to be able to counter denial of service (DoS) and other attacks from the network. There are many security features (i.e., Firewall) inherent in today's Ethernet switch architecture to allow the network to deal with any internal or external attacks.

The Ethernet switch in the IP phone also needs to support at least some security features, as the IP phone can also be a recipient or an originator of such an attack. Features such as MAC address locking and Ethernet port locking in the IP phone enhance the network's ability to deal with attacks quickly and efficiently. A host of other features, including MAC table freezing, port authentication (802.1X), spanning tree protocol (STP), classification engines (classifying packets using packet information from layers 2/3/4) and virtual local-area networks (VLANs) (802.1Q) make sure the IP phone poses no threat to network security.

Encryption algorithms that run on the central processing unit (CPU) or independently on a dedicated hardware accelerator give IP phone vendors a further layer of security, enabling the encryption of incoming and outgoing calls as well as any of the data that runs through the device. This becomes vital when vast quantities of highly confidential information are being exchanged over "unsafe" connections between corporate and remote sites or even between vendors and clients that do not share networks. Standards such as advanced encryption standard (AES), data encryption standard (DES) and 3DES are found most often in a state-of-the-art IP phone with 128-bit encryption additionally supporting the hashing algorithms secure hash algorithm 1 (SHA-1) and Message-Digest algorithm 5 (MD5), as well as a true random number generator (TRNG).

WLAN, Bluetooth, or DECT Support
Wireless technology is pervasive and influences everyone's lives. IP phones can and will take advantage of this development. Bluetooth headsets are now a standard feature on most mid- to high-end IP phones and are no longer an "exotic" extra.

The widespread deployment of wireless local-area network (WLAN) (802.11x) access points in the enterprise will give rise to more WLAN, or Wi-Fi, phones. This is a challenge for the fixed-line IP phone. To date, however, this has not been proven to be true, as WLAN phones struggle to achieve the high level of quality of service (QoS) in a network phone as well as compete with mobile phones that work within and outside the enterprise.

DECT is extremely robust, low-cost, and stable with strong benchmarks in standby times and reach. Adding VoIP to this makes it a very interesting argument for making IP phones mobile or integrating a base station into the phone to work with existing handsets.

Summary
In the past, IP phone penetration in the corporate network was hindered by two main factors: low voice quality and a high price tag. The former issue has long since been solved with the advent of advanced voice codecs. With respect to cost, VoIP installations are now more than recouping their owners' investments with savings in the connection of remote offices, the costs of moves and replacements, and the improved user productivity employing a wide range of VoIP applications.

Next-generation IP phones will become an integral part of the GE network. Excellent voice quality will become the norm, and IP phones will support a wide range of industry-standard and customer-proprietary voice codecs as well as other audio-related features, including MP3, stereo, and polyphonic ringing.

Educational content provided by Dr. Carsten Sydow and Daniel Hartnett, Infineon Technologies AG

bar