It takes a lot more than a high-quality handset to make a high-quality phone call. Unlike wireline phone service, mobile devices rely on the often unpredictable, over-the-air radio transmissions to carry time-sensitive voice traffic. Add to this the transforming of voice calls into IP-routed traffic, microphones and speakers disconnected from the device via Bluetooth, high-bandwidth competition from e-mail, data and video downloads, and multiple hand-offs between network routers and gateways, and you have a lot of places where voice quality can get degraded.

By Charlie Baker

Convergence also is piling on complexity to the mix. For example, Comcast and Sprint's new "quad-play" Pivot service links your mobile phone to your home broadband and cable TV services, taxing bandwidth even more.

To address consumers' voice quality concerns, some operators are promoting quality based on the percentage of dropped calls, network reliability and the speed of their wireless cloud. But what operators really want to do is deliver the highest quality of experience (QoE) to a subscriber, no matter how that customer is accessing the network. Although most operators have systems in place for measuring things such as a handset's RF signal strength, it's not enough information to identify a host of other potentially damaging quality issues.

Work is already under way to migrate infrastructure to higher bandwidth and higher speed connections to the base stations. Carriers are starting to create and enforce service level agreements (SLAs) between network providers and network peers to measure the customer experience across the network. The goal is to proactively identify performance degradations before they impact users, satisfy contractual SLAs, and cost-effectively troubleshoot user-impacting problems when they do occur.

Consequently, voice service providers that take quality seriously are employing testing strategies via  converged service assurance solutions that address the user's QoE two different ways: through the use of active (on-demand) testing and passive (live) monitoring of traffic.

Active testing injects traffic between network segments to isolate disruptions before they affect the user's perceived voice quality, while passive monitoring assesses voice traffic across the network to discover quality trends before the problems become widespread.

The mobile handset is key to understanding voice quality issues in both testing approaches. One passive monitoring approach is to deploy software agents on handsets to measure quality using measurement standards like a Mean Opinion Score (MOS), R-factor, echo and jitter. But, agent technology requires handset manufacturers to install the operator's software on every device they ship. This may be good for the operator, but it can be a pain for the manufacturer and thus a contract stumbling block.

A simpler passive alternative is for the handset manufacturer to adopt standard protocols, such as Request for Comments (RFC) 3611, RTP Control Protocol Extended Reports (RTCP-XR), which provides a way to measure and manage voice performance by reporting detailed RTP and VoIP metrics.

Operators can combine RTCP-XR with the SIP Service Quality Reporting Event, a draft standard specification popularly called the Johnson draft spec from the Internet Engineering Task Force (IETF). With these standards, operators can define a method for retrieving voice quality information at the end of every user's call. The typical approach is to have a SIP message send quality metrics at the conclusion of a call to a service assurance system to correlate information across multiple points in the network and generate targeted reports on voice quality. Although few operators are using SIP all the way to handsets today, they definitely will as they migrate toward IP Multimedia Subsystem (IMS) based networks.

The emerging picocell/femtocell market is adding more complexity to wireless infrastructure. Picocells are wireless communication devices that cover a small area, such as the home, through which mobile phone users can access the operator's network wirelessly through their home broadband connection rather than over-the-air to the cell tower. It's similar to a Wi-Fi access point, but for mobile handsets.

Picocells will become another point in the network where operators will need to know about quality. Using the same passive agent or protocol-based approaches will enable operators to extend quality testing to another point in the network where their customers may experience voice quality issues.

As with passive monitoring, standards are key enabling technologies for active (on-demand) testing. With active testing, an operator triggers a transaction to the user's handset or picocell – completely undetectable by the user – to measure quality. Standards used by the operator and handset manufacturer include RTCP-XR, and the IETF drafts for SIP Media Loopback and the 2-Way Active Measurement (TWAMP) protocol.  With this approach, the operator usually deploys a service assurance appliance in the middle of its network. This appliance initiates a call to the customer's device, with protocols looping the payload back to the appliance so that it can be measured.

Active testing enables operators to test devices across the network 24/7, and identifies potential problems before customers detect them and either call for assistance or return their equipment to the store.

Clearly, not everyone is ready to drop their landlines.Yet. But, with great service assurance coverage, that trend may soon accelerate. Combining active testing and passive monitoring provides carriers with visibility to service performance across their network, resulting in reduced truck rolls and tighter SLA enforcement, and gives consumers higher quality calls every time they press send.

Baker is director, product management at Chelmsford, Mass.–based Brix Networks.