|
Voice over IP (VoIP) is an application deployed over converged networks where service impairments can be difficult to identify and resolve. Lessons learned from current VoIP deployments show that a gap in the service assurance process can result in excessive expenditure of time and resources involving many technical disciplines to resolve service issues. With the rapid growth of VoIP deployments, a resource-intensive approach can quickly overwhelm an organization. Implementing the following items will result in a scalable and cost-effective program for VoIP test and service assurance.
1. Key Performance Indicators (KPIs) � The VoIP product/service performance requirements should be mapped to network Key Performance Indicators. The KPIs will aid in identifying relevant network measurements and performance alarm thresholds. Common KPIs include time to dial tone, post-dial delay, call success rate, dropped calls, Mean Opinion Score (MOS), packet loss, latency, jitter, and echo.
2. Pre-deployment Network Qualification � Exercise the network in advance of service rollout with a combination of physical loop tests, call generators, probes, and analyzers to determine if the network can reliably deliver VoIP service with the required Quality of Service. DSL testing involves copper loop pre-qualification to detect load coils, bridged taps, wet sections, grounds, etc. Coaxial cable testing will ensure that the forward and return paths meet DOCSIS requirements. Automated call generators, simulating subscriber calling patterns, will indicate if the network is correctly configured (softswitch, dial plans, routes, policies, firewalls, PSTN gateways) and has bandwidth capacity for VoIP service.
3. Installation Testing � A large percentage of VoIP service impairments are caused by inside wiring, customer handsets, or customer premises equipment configuration. For professional installs by company technicians, service at the residential gateway RJ-11 port should be tested to validate dial tone with calls to a remote analyzer or handheld test set. Phone jacks in every room should be tested to fix inside wiring problems. Customer self-install business models can be validated with automated loopback testing triggered when the residential gateway registers with the network. The self-install manual might also instruct the customer to place a call to an automated response test system.
4. Loopback Testing � Automated testing of VoIP services through scheduled loopback test calls to the residential gateway is a proactive means for detecting signaling and media performance impairments. This has the advantage of finding faults before the customer goes off-hook and notices a problem. One strategy that balances test coverage without overloading the network is to test a small percentage of the subscriber population chosen at random every 24 hours. Additionally, some residential gateways allow loopback testing concurrent with a normal subscriber call; the loopback call is treated as a third session on the same line. This feature offers the possibility of interactive loopback testing by a customer service representative while speaking with the subscriber to troubleshoot a complaint.
5. Signaling Analysis � A signal analyzer provides call trace and protocol analysis to troubleshoot call setup and teardown issues for VoIP and PSTN calls. Historical call records enable investigation of customer complaints after-the-fact by Tier 2 or 3 support personnel. Newer initiatives seek to combine call flow traces with VoIP bearer channel (RTP media) Quality of Service measurements for a complete view of the customer experience; because signaling messages alone do not indicate speech quality. Call quality metrics can be collected from RTCP XR (RFC 3611) data in a softswitch Call Detail Record (CDR) database, passive monitoring probes, and active test systems. The greatest network coverage is found in the CDR database.
6. Trouble Signatures and Corrective Procedures � Troubleshooting a VoIP call can be a complex undertaking involving experts from many disciplines to solve the puzzle, even when the impairment is readily apparent and persistent. The development of trouble signatures, probable causes, diagnostics, and corrective procedures is needed for operational effectiveness. The result is a comprehensive troubleshooting guide for the organization and system requirements for service assurance and test systems.
7. Event Correlation � A major challenge is to integrate the telemetry from numerous network elements in the VoIP service delivery path � CPE, access, edge, and core networks, PSTN interface, service platforms, and test systems � into a meaningful view of service-affecting conditions to determine root cause. A layered architecture is necessary to filter information from the network elements and test systems before forwarding to higher level manager-of-managers for expert system identification of trouble signatures and root cause.
8. Automation � VoIP networks are highly dynamic due to rapid subscriber growth and the inherent flexibility of IP networks. Bottlenecks do occur and the network can be quickly overwhelmed with systemic failures, resulting in an avalanche effect when a capacity threshold is crossed. A high degree of automation is needed to identify trouble signatures, enact corrective measures and be proactive with customer notifications to reduce the volume of trouble reports. Test and monitoring systems play a key role by providing benchmark references for early detection of trends, threshold crossings, and alarm notifications.
9. Training � Training touches all VoIP service provider staff, from field technicians to network operations to customer service representatives. VoIP technologies require updating of knowledge and skills as network intelligence is increasingly pushed out to the residential gateways (e.g., per-subscriber VLANs). Training programs for field technicians and call center personnel, in conjunction with the tools recommended in item #6, can help optimize their ability to resolve VoIP issues. Call center personnel can be especially effective in the early detection of unusual conditions.
10. Configuration Change Management � Rigorous change management processes should be followed for the carefully controlled introduction of approved updates to the network. What worked for managing high-speed data (Internet) services can have unforeseen consequences for VoIP service. Lab networks do not fully duplicate the environmental conditions of large-scale production networks, making it difficult to fully validate a change before deployment. Risk can be mitigated by testing new configurations against in-network VoIP test probes that emulate subscribers making and receiving calls, as a confidence check. IT
Robert Kinder is business development manager at Tollgrade Communications (News - Alert). For more information, please visit www.tollgrade.com.
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 [email protected] or by phone at 800-290-5460.
|
SUBSCRIPTIONS
TMCnet LOGIN
WEBINARS
