Automated and manual testing approaches for VoIP performance

VoIP PERFORMANCE TESTING TOOLS

"There will always be a need for manual testing"

Terry Caterisano (quoted in Linask).

VoIP Services

Conducting automated and manual tests on Voice over Internet Protocol (VoIP) can help VoIP service providers circumvent expensive and embarrassing problems as well as support their commitment to ensure their network delivers acceptable call quality and non-stop availability. With some VoIP services, one may only call other individuals using the same service. Other VoIP services, albeit, may permit one to phone any local, long distance, mobile, and/or international numbers. While various VoIP services work only over a computer or a particular VoIP phone, a number of other services permit a person to make calls with a traditional phone connected to a VoIP adapter. Despite differences, however, VoIP service providers and vendors reportedly routinely test VoIP performance to enhance its performance (Voice-Over-Internet Protocol; Linask; VoIP Performance Testing...).

During the research paper, the writer presents a sampling of several technical and business sources to relate relevant reflections regarding VoIP performance tools. The writer also asserts, as the quote at the start of this paper indicates, that in testing, the person, more than any system in both automatic and manual processes, proves to be the most valuable component.

2. Does the VoIP application affect the performance of other applications?

Voice, by definition, exists as analogue in nature, while voice quality proves highly subjective. One network may deliver a number of different perceptions of quality for various voice types. When measuring voice quality, tests involve two essential advances, analogue and digital. An analogue voice quality tester produces a real voice call across a network and weighs it against the received call with the reference call. A digital voice quality tester analyzes the packet data stream containing the VoIP data and measures packet loss, packet delay and packet jitter. The tester then uses the ensuring information I to calculate the likely voice quality, expressed as predictive Mean Opinion Score (MOS) (Voice-over-IP (VoIP) and network performance).

When the VoIP voice quality proves inferior, the service provider, vendor or customer may need to test cabling channels as over time, moving, adding and/or changing cables can contribute to their degradation. Circuits may also contribute to substandard voice and call quality. Other quality challenges may require a remote monitoring (RMON) tool to track VoIP performance. RMON, a network management term constitutes a "specification for collecting and reporting advanced network activity data. Data collected using RMON can be used to analyze remote network performance, topology, faults, departmental usage (for accounting) and protocols" (RMON, RMON Section). The Internet Engineering Task Force (IETF) developed RMON to support monitoring and protocol analysis of Ethernet and Token- Ring LANs. RMON depicts the industry standard specification that furnishes to a great extent of the functionality contemporary proprietary network analyzers and protocol analyzers offer.

According to Linask, Caterisano stresses that no standards for testing VoIP currently exist. The majority of provide a level of automation and remote control over VoIP devices, but only for basic tasks, like loading a script, initiating its start, stopping the script, and capturing results. No consistency, albeit, exists between vendors regarding how they facilitate those capabilities. Even though the need for manual testing will always exist, automation can offload 50% of the work the tester may perform daily onto a structure that runs tests overnight and on weekends.

When a network server or vendor plans for VoIP, the planner should expect the peak in use to occur all the time. If peak service scenarios are not pre-planned, when heavy processing does occur, the provider or vendor's voice signal will degrade. In addition, examining buffer space and discards on active electronics proves vital. As a function, switches discard packets. When the buffers exceed their full capacity, they drop packets and require the sender retransmit, an adverse feature for voice. Even though one may set up VLANs and priority, overloaded gear will prove futile. Instead, one will need to check discards on any uplink port and any port generally attached (for example in the IP switch) (VoIP performance testing fundamentals).

In addition, the tester may discover a number of errors in SNMP data that also merit investigation. Bit errors, which may be perceived as InErrors and OutErrors prove most significant. Some manageable systems will not permit one to drill down further into the error state, albeit, some systems allow this action and consequently enhance the troubleshooting process. When these errors surface, one should initially test the cabling channel connected on that port as well as confirm the installation of the latest revision of software on the system and the updated, current status of the firmware calibration. The tester also needs to ensure interfaces and/or patch cords are relatively new. As each implements a limited number of mating cycles, a specific channel may appear to be malfunction when it is performing correctly (VoIP performance testing fundamentals).

The tester also needs to examine duplex configurations as having full duplex links duplex proves vital to performance. Mismatches and/or channels auto-negotiated to half duplex will further limit VoIP operations. Half duplex links may result from a hard setting in either the switch or at the workstation as well as faulty cabling, including channels exceeding the maximum distance designated for VoIP performance. When the tester corrects errors, he will need to schedule a following network pulse in 30 days. This time element allows for things as month-end processing and other functions that do not occur daily (VoIP performance testing fundamentals).

In the article, "Operation: VoIP: Testing and Monitoring Are Complex. But Key for Solid

Results," Craig Kuhl explains that VoIP testing includes the following four stages/segments

1. Research and Development test equipment

2. Manufacturing test equipment

3. Installation and maintenance test equipment

4. Monitoring test equipment (Kuhl, ¶12).

Kull asserts that the ultimate criteria to ensure the peak VoIP Quality of Service (QoS), reliability and performance proves challenging. As VoIP constitutes a new venture for the cable industry, it mandates fresh skill sets, tools, equipment as well as innovative corporate engineering and perceptions. In response to the Yankee Group's projection that the growing VoIP business would reach 18 million subscribers by 2009, the vendor community designed and administered a number of crucial tests to pass or fail an IP-enabled voice network. Cable VoIP, like broadband VoIP, comprises a legitimate business. This contributes to the expectation that competition between service providers will significantly increase during the next 10 years. Critical tests, albeit, will need to continue to help ensure efficient service. As this growth occurs, service providers and vendors will need to implement even more test and monitoring equipment (tools), particularly in areas like troubleshooting (Kuhl).

As VoIP customers expect to receive excellent service, both cable operators and broadband service providers must be more pro-active in testing and monitoring VoIP performance. VoIP networks have their sensitive side, too. Therein lies the test and monitoring challenge. Randal Burke, senior director for national communications engineering and operations for Comcast Corporation, states that managing VoIP does not depict typical network management. Burke stresses that cable needs to implement a more active view of nodes prior to their impact on customers (Kuhl). Cable and broadband service providers and venders must maintain quality test and monitoring systems. They must track performance issues such as jitter and latency as well as regularly address multiple VoIP gateways.

Kuhl projects that VoIP testing strategies will solidify as VoIP reaches its peak deployment period by 2010. Resolving weaknesses in VoIP requires that the tester first identifies the weaknesses and relating them to baseline elements of packet configurations and root causes. In the realm of cable VoIP, the reportedly most powerful tool for testing VoIP, according to Kuhl, consists of reactive testing. This permits the service provider to extract actual call-in data, enter it into their database and trace the root cause. This permits the tester to view each call and secure a clear picture of the network's performance.

Conclusion

Amidst progress in VoIP monitoring and testing, the quest continues for the ideal testing system and its support equipment. Testing proves complex as locating and employing techs with the skills needed to multi-task within the IP worlds, who speak the VoIP data and telecom languages presents even more challenges. Meeting contemporary testing challenges, nevertheless, and conducting automated and manual tests on VoIP, as noted at this paper's start, can help VoIP service providers avoid expensive and embarrassing problems. It can also enhance their commitment to ensure their network delivers acceptable call quality and non-stop availability. Fulfilling this commitment will help ensure customers remain loyal in the midst of contemporary competition. After all, in providing VoIP business, the customer, as the individual in the testing process, more than any system in both automatic and manual processes, proves to be the most valuable component.