VoIP has become a household word. Cable MSOs and Telcos are offering VoIP home phone service to a very receptive public seeking alternatives to their higher-priced, inflexible plain old telephone service (POTS). But paying subscribers sign up with expectations providers need to offer wireline quality and reliability, efficient installation, and competitive pricing to differentiate their services from free, Internet-based voice messaging applications and low-cost cellular plans.
Some providers are making the grade a recent international call quality study shows that Cable VoIP using low-compression CODECs actually sounds better and connects calls faster than traditional PSTN (define - news -alerts) service. As consumers spread the word about VoIP as a cost-effective, high-quality alternative to traditional phone service, service providers will experience ever-stronger demand.
Maintaining service quality during large-scale deployment, sometimes to millions of subscribers, is a key challenge operators must tackle to capitalize on VoIPs early success. As high-priority, real-time voice packets begin to fill networks, quality must be monitored and optimized or service quality will quickly deteriorate.
Successful service providers plan large-scale VoIP deployment using a comprehensive, service-quality focused strategy. A recent Business Week article (Nov. 28, 2005) outlined how Comcasts rapidly growing VoIP service maintains PSTN-quality using unique provisioning and monitoring testing. Comcasts test strategy allows them to offer a 30-day satisfaction guarantee to new subscribers an approach that has led to rapid customer growth. Operators who leave service quality to chance have been less fortunate, experiencing a high rate of customer churn as subscribers return to trusted incumbent providers. In a highly competitive industry, quality is a proven differentiator and a key asset to survival.
Actively Testing VoIP
To ensure effective VoIP deployment and continuing service quality, providers rely on passive and active test solutions for provisioning, monitoring and service level fault management. Passive and active telephony testing techniques were first developed by carriers over 20 years ago to measure the quality of digital PSTN voice networks. In traditional telephony, passive test systems are used to measure signaling performance, while active bit-error-rate (BER) testing is used to monitor PSTN call quality. The same techniques have their analogs in the VoIP world. Passive testing involves either collecting traffic statistics from softswitches, media gateways, and other network elements, or requires placing test probes at strategic network locations to sniff live IP traffic and filter out VoIP packets. A variety of algorithms examine packet header and payload information to calculate call signaling performance and some basic quality metrics like MOS, packet loss, jitter, and latency. Some passive solutions
also report the breakdown of coexisting email, web, peer-to-peer, file transfer, and multimedia traffic. Passive testing is commonly used for basic service quality monitoring, and to assess network usage for capacity planning, performance and routing optimization.
Passive testing solutions have known limitations, some of which are specific to VoIP applications. Since passive testing relies on measuring live VoIP traffic, it cannot be used for provisioning or pre-deployment testing, necessary to ensure customers expectations and service level agreements will be met after installation. Security concerns around VoIP have resulted in increased packet-level encryption, rendering them invisible to passive test systems. Adding to the problem, probes can only analyze traffic on an operators IP network, providing limited visibility into the majority of calls that take hybrid routes over partners PSTN, IP, and mobile networks. Since call quality is affected by analog/IP conversion (CODECs), echo cancellers, customer premises equipment, wiring, firewalls, encryption algorithms, and network policies outside the reach of passive probes, measurements from these systems cannot accurately represent user-perceived end-to-end call quality.
As a result of these limitations, VoIP providers often follow the example of PSTN operators and use a combination of active and passive testing techniques for monitoring and fault management. Whereas passive testing offers good visibility into call signaling performance, active testing offers operators the best possible technique to analyze end-to-end media quality, the true end-to-end customer experience of service quality. In active testing, probes place short test calls that are recorded then analyzed by industry-standard algorithms to produce a detailed, ear-to-mouth service quality assessment. Active testing can be used with any standard telephony interface including analog (2 wire), ISDN, mobile, or IP. Test calls can be dialed using SIP, H.323, MGCP, PRI, SS7, CAS, or mobile signaling standards. Because active testing is not limited to IP networks, it can be used to test calls placed over hybrid routes for example from a cellular, through a PSTN network, to a VoIP phone measuring end-to-end
speech quality, call setup and session information. Active testing works well for provisioning and pre-deployment audits where test calls are placed alongside existing or simulated network traffic. Using automation, test plans can be scheduled to execute at regular intervals for trending, monitoring, and service level fault management. On-demand test calls can be used for remote troubleshooting.
Active Test Layouts
Test solutions are primarily designed to help operations staff and technicians find and fix problems before they are experienced by customers like radar or night-vision binoculars, they help to identify, visualize, and isolate problems that would otherwise be impossible to detect. But just as you wouldnt use binoculars to read a book, different departments need different tools to address their specific test requirements. A technician installing and troubleshooting VoIP in the field needs a portable, straightforward instrument that can quickly assess service quality from a users perspective; operations personnel need tools to remotely identify and replicate service issues to speed service restoration and minimize technician dispatches. Operations require service-level fault management. Billing needs to validate and maintain service level agreements (SLAs) with straightforward reporting. Active testing addresses these various requirements using different test layouts, optimized for provisioning,
troubleshooting, and monitoring. Test layouts include innovative test methods that can be combined to provide comprehensive coverage from subscriber to central office/head-end, and all points in between.
Single-Ended Test Layout
The single-end (ping-style) test layout measures voice, signaling, fax, Internet, and modem service quality without using a far-end probe. A test probe records call progress and content for playback and analysis. A location database provides public telephone, fax, modem, or IP destinations for testing on or off the service providers network. Tests calls are designed to assess service quality quickly minimizing long-distance billing charges by hanging up just after the connection is established. The test server typically reports results by area code, city, network provider or country. Single-ended testing is typically used by long-distance service providers to verify call quality over partner carrier networks, and to automate least cost routing processes. It can also be used for troubleshooting to specific phone numbers, and to routinely spot-check quality on geographically large networks.
Probe-to-Probe Test Layout
In probe-to-probe testing, test calls are placed from one probe to another, providing complete control over test stream content and quality measurements in both directions. Tests can be initiated and received on any standard signaling interface (analog, PRI, SS7, mobile, SIP, etc.) over NextGen and legacy networks. Tests record call progress as well as the actual voice/data traffic for playback and signal analysis. Probe-to-probe testing is typically used to monitor VoIP, PSTN, video, Internet, and fax/modem service quality over large-scale IP, TDM, and mobile networks. Large providers will often place several dozen probes in POPs and central offices. Test automation is used to schedule network-wide test calling patterns between all probes in the network. When a service quality issue is identified, the operator can choose to focus test calls over the problematic route to quickly isolate and troubleshoot the problem.
Responder Test Layout
Responders replace far-end test probes when portability and cost are primary concerns (Figure 3). Responders are ideal tools for field technicians performing day-of-install testing and troubleshooting, and for regular customer-site SLA validation. Active test responders can receive and transmit test traffic, and are commonly connected to either 2-wire analog or IP ports. Analog connections allow technicians to replace a telephone handset with a responder to measure user-perceived quality. Residential VoIP installation usually involves using an analog responder to quickly test call quality from all the phone jacks in a home, measuring MOS, distortion, echo, and call connection statistics, packet loss, round trip delay and DTMF tone transmission transparency to ensure the customer will be satisfied with the new service. IP responders are usually used to test RTP traffic streams for VoIP and video quality assessment.
Loopback Test Layout
Integration with softswitches and network components allows active test systems to conduct analog and RTP/IP loopback tests from test probes to CATV transponders and subscribers multimedia terminal adapters (MTAs). Loopback tests (Figure 4) reduce the need to deploy technicians to resolve network and service performance issues, and require no investment in portable test equipment. Loopback testing is often used by Cable MSOs to streamline operations and reduce support costs, and to accelerate problem resolution. Loopback testing can quickly isolate a service issue by testing to a subscribers home, neighboring homes, and nearby transponders. Data from RF-level transponder tests, IP, and analog loopback tests to the MTA can be correlated to remotely determine on what network layer the problem originates and what steps are required to correct the problem.
Interactive Test Agent Layout
Voice-guided testing is the most recent active test technology; the test probe hosts an interactive voice response agent that uses natural speech to guide technicians or subscribers through a series of simple steps leading to detailed speech quality, DTMF, echo and noise tests without any far-end responder (Figure 5). Designed to reduce installation, troubleshooting, and trouble ticket resolution time, interactive testing is ideally integrated into customer support systems to provide immediate test data when customers call with service issues. Test results can be used by customer service to guide the user through corrective actions, by operations staff to remotely correct the problem or to dispatch a field technician if field repair is required.
OEM Tester Layout
Some active test systems test from a probe to third party test sets, TDM responders and network elements as test endpoints. This functionality enhances the flexibility of the test system by leveraging existing capital investments. In some cases, active test systems can add VoIP, Video, and other service level functionality to physical and network-layer DSL, Cable, and FTTx testsets that would otherwise have to be upgraded or replaced. Tests can be initiated by the test server, the legacy operational support system, or by the handheld testset if they support integration.
The Big Picture
A test automation platform is used to manage and integrate the results from the different test layouts and applications, centralizing all aspects of VoIP service delivery and maintenance. An integrated test platform enforces centralized standards and test plans, provides event correlation and reporting, and is easily integrated into existing operational support systems allowing all departments to work and communicate within a single, quality-focused framework. With a single, integrated test platform, departments can efficiently collaborate to deliver the best possible VoIP service. For example, a technician troubleshooting VoIP at a subscribers home would also have access to historic day-of-install test results, and NOC personnel can run remote tests to neighboring homes and share the data with the technician, effectively isolating the problem in time and space to accelerate problem resolution.
Despite the challenges faced by service providers deploying VoIP, it is possible to rapidly scale subscribers while maintain PSTN-grade service quality. An integrated, comprehensive test strategy that encompasses VoIP delivery from day-of-install to proactive service monitoring can be addressed using automated, active test systems. An effective service level test automation platform can integrate results from different test layouts, allowing all departments to work towards their respective goals within a common framework. With a test platform, a service quality focus can be integrated into existing operational support systems, fault management systems, and existing business processes giving the visibility required to compete in todays competitive telephony landscape. IT
Scott Sumner is director of marketing at Minacom, a provider of Service Level Test Automation solutions for Telcos and Cable MSOs. For more information, please visit the company online at
www.minacom.com (news - alerts).
If you are interested in purchasing reprints of this article (in either print or PDF format), please visit Reprint Management Services online at www.reprintbuyer.com or contact a representative via e-mail at firstname.lastname@example.org or by phone at 800-290-5460.