May 2004

Wireless VoIP - Other Voices

BY RICHARD WATSON

In earlier articles, the discussion has focused on wireless VoIP that was telephony orientated, with full duplex capability and enhanced telephony features. In the world of “voice,” however, this is not the only value-added feature that can utilize audio transmissions across a network fabric. As VoIP and WLAN technologies mature, there are more variants of audio applications emerging, offering different application functionality than standard telephony services. Since these applications rely on being able to send a voice packet stream over the network, they still depend upon high-quality transmission of voice from a mobile terminal to some application entity within the network, but must meet a more diverse application requirement. This article looks at some of these “other voices” and how they are technically addressed and bring greater value to various segments of the enterprise market.

PUSH-TO-TALK (WALKIE-TALKIE)
In many enterprises, associates make use of “Walkie-Talkies” — dedicated VHF-based devices used to communicate with multiple team members simultaneously. This mode is half-duplex where only one person within a group can speak while all other members are in a listen mode. When the initiator finishes speaking, another member of the group may initiate a response where all can listen to the new speaker. Because operation of this device requires pressing a button while speaking, it is also known as “Push-to-Talk.”

As traditionally implemented in commercial products, the Walkie-Talkies have been devices that support this single function. Depending upon the scope of coverage and number of service channels required, these devices can be quite expensive. With the deployment of WiFi within an enterprise, the question arises: “can this function be supported over WiFi?” The answer is “YES.”

There are new products emerging in the WiFi market that support features labeled as “Walkie-Talkie” or “Intercom” that replicate the traditional half-duplex functionality and even go beyond this functionality. The architecture for these products is based on support of audio packets being transmitted via multicast addressing. This architecture is straightforward and aligned with basic network design fundamentals; however, there are some wireless hurdles to be addressed.

Using a MAC layer multicast to carry real-time audio data poses a problem in a standard WiFi (802.11) product. As originally architected, a WiFi Access Point must transmit any multicast or broadcast frames immediately if there are no mobile clients in Power Save Polling (PSP) mode. However, in an operational wireless LAN, there is a high probability that at least one client will be in PSP mode and the frames will be queued and transmitted on a Delivery Traffic Indication Map (DTIM) frame (a frame that announces to all associated devices that there is data queued) interval. On the DTIM interval expiration, all queued multicast/broadcast frames are then transmitted, which may sound okay, but the default DTIM interval for some infrastructure products can be as much as one second. This means that any inbound multicast frames will be queued and transmitted on a one-second interval, which results in choppy, totally unacceptable voice quality. Now that vendors are supporting multicast-based applications, how can good voice quality be achieved? In the current marketplace, there are two approaches to creating an environment where such applications will work well.

The simplest solution involves a minor configuration change of the wireless network to shorten the DTIM interval setting. In fact, many of the commercially available products recommend this approach, which can reduce the DTIM interval to less than 100 milliseconds (msec). While this produces the desired effect in improving the voice quality with multicast applications, there is a downside. The downside is that all associated devices must respond to the rapid DTIM signaling, causing a severe impact on the observed battery life of each device. With this approach, network managers are left with a quandary: optimize the voice application or device battery life?

A more elegant solution involves enhancing the functionality of the Access Points to provide special handling for multicast traffic. In this case, the Access Point must be aware of application requirements based on packet addressing values. Some currently available WiFi products allow administrators to specify a multicast MAC address to identify the “Walkie-Talkie” application audio frame. Once identified, the AP will apply special processing to these frames causing them to bypass queuing and simply scheduling it for immediate transmission. Other non-specified multicast and broadcast frames will be handled in the standard manner based on DTIM interval transmission. The assumption is that the voice-active clients will not be in PSP mode and will receive the frame while any PSP clients will remain in that state undisturbed. This approach is the best compromise to support good voice quality for the multicast applications, while also optimizing battery life for associated client devices. However, without a standards-based specification for this feature, the implementation is proprietary. Thus, it is important to inspect the vendor support feature set if you desire to implement this kind of application.

ALTERNATE VOICE APPLICATIONS
For certain vertical markets there are industry specific applications that need to take advantage of the underlying telephony (PBX) functionality. These applications meet industry-specific data functionality, but also place a requirement on being able to support some kind of voice connectivity.

One example of such an application is customer relationship management (CRM) where a user may wish to talk to a customer and desires to simply select a name within an application window to initiate a phone call. This reduces the operation to a single task rather than requiring the selection of a “softphone” application window and explicitly entering the customer’s phone number. This creates a real productivity advantage.

Another example of such industry specific application is the Nurse Call application. This application has a more complex configuration requirement, but still involves the use of a full-duplex phone call to fulfill the functional requirements. With a Nurse Call application, a hospital patient signals for assistance by pressing a button at the bedside. This alert is processed by the central Nurse Call system, which then signals all the active ward nurses, via text messaging and asks if they are available to respond. Depending upon the nurse’s availability, they respond to the text page and the system automatically sets up a phone call between the nurse and the requesting patient. With such an application, the nurse’s interaction with the system is minimized, while maximizing response times to the patient need. Again, a significant productivity advantage is created in an industry where there is an extreme shortage of nursing staff.

The underlying technology requirement for these applications remains reliable wireless VoIP support. As various market segments adopt VoIP and WLAN technologies, there will be additional kinds of “alternate” voice applications.

VOICE RECOGNITION
A more familiar telephony feature is voice recognition, most frequently encountered with Interactive Voice Response (IVR) systems, often encountered when calling into large corporate support centers, banks or a doctor’s office. This technology is not traditional telephony, but still requires the generation of a reliable audio packet stream. Implemented as a “phone call” between an active user and the voice recognition engine, a voice “session” must be established where the user is prompted by pre-recorded auditory instructions and the user responds by speaking the appropriate response. This response is then processed by the speech recognition engine, which then acts upon the information provided to fulfill the desired operation. Simple voice recognition can be used to navigate menu options for service or support systems (e.g., “press ‘1’ or say ‘one’”). More complex uses of this technology can be used for phone directory lookup or for user authentication via voice print matching.

OTHER AUDIO APPLICATIONS
Other applications are possible where a user dialog is recorded by archiving the RTP packet stream, which can be useful in a number of industries. For example, call centers typically record conversations “for training purposes.” Doctors traditionally make patient notes verbally and later transcribe them to hardcopy. These applications can be accomplished through the use of an Ethernet VoIP device, but a mobile WiFi appliance enhances the usability and user productivity for these applications.
The potential of developing wireless devices with voice capability has yet to be exhausted. As new “voice enabled” WiFi devices come on the market, other applications will be developed to meet expanding needs and provide productivity to a variety of user communities.

SUMMARY
As the WiFi market continues to expand, more “voice enabled” devices will enter the market to fill an ever-expanding set of application needs. Basic wireless telephony addresses a large segment in the general market, but industry specific applications are emerging that rely on the same basic primitive set of services – reliable, secure audio stream transport. Management of this broad mix of telephony resources may become a real challenge, especially when all the applications attempt to use these services at once. In this case, configuration management, application design, and user training will need to be addressed to help control the user experience. All in all, this evolving technology is bringing a richer set of features and functions to the end-user community.

Richard Watson is director of telephony product marketing for Symbol Technologies’ Wireless Systems Division in San Jose, CA. Prior to taking on the marketing role for Symbol’s NetVision family of WiFi Telephony products, he managed the software engineering team for three years and was responsible for developing Symbol’s WiFi Telephony products.

If you are interested in purchasing reprints of this article (in either print or HTML format), please visit Reprint Management Services online or contact a representative via e-mail at reprints@tmcnet.com or by phone at 800-290-5460.

[ Return To The May 2004 Table Of Contents ]