Successful migration to IP telephony is a Management Issue

Quality of Service without Compromises
Voice over IP (VoIP) technology promises substantial cost savings thanks to the convergence of voice and data transmission. To reap these cost savings, the existing IT infrastructure must be suitably configured and managed. A prerequisite is having meaningful performance data that allows IT administrators to work in a target-oriented manner throughout all VoIP migration phases, including planning, installation, testing, operation, and expansion of the network.

Convergence has caught on for two reasons: it increases employee productivity and decreases voice / data transmission operating costs. Productivity goes up especially thanks to new multimedia applications and the intelligent integration of telephony in IT usage. Costs go down when redundant networks become unnecessary and the administration of IT infrastructures becomes more efficient. Convergent technology also helps to further decrease costs because it makes moves, expansion, and changes easier to execute.
An additional advantage of this new trend is that a company with several locations can present itself as a cohesive unit when on the phone with a customer. Various phone numbers for different locations, separate offices, and field staff are no longer required. The future belongs to uniform “addresses” for voice and data communications, without having to give up already known phone, fax, and company numbers.

Convergence really came in vogue after individuals and companies had experimented with it fairly intensively over the last ten years. Its present emergence can be attributed to the fact that past technology was not quite able to guarantee the required quality in voice transmission over the data networks. Sporadic lapses, unintelligible portions, and complete disruptions of the connection were some of the more annoying consequences – and thoroughly unacceptable in the business world.

The situation has changed fundamentally in the meantime. Today, not only are higher bandwidths available at more cost-effective prices, but so are innovative VoIP processes and sophisticated management technologies. Using these, administrators can assure the appropriate voice quality of their company networks throughout numerous locations. For example, to guarantee quality of service (QoS), they can use network components with integrated functions that reliably identify the IP-telephony traffic and allocate it the best possible performance and quality within the network.

Management from start to finish

Such features are particularly effective when voice-compression technologies are selectively utilized in wide area networks (WAN). They significantly expand the bandwidth for all applications, and thus automatically improve the voice transmission quality of VoIP applications, too. Their quality depends on several factors: bandwidth, uninterrupted availability of the IP network, response time behavior, variances in transmission time, and the error-free interaction of network components – all of which fall under network performance. This performance can only be ensured by consistent network management and systematic control of the IP traffic.

In contrast to file transfers, messaging, and other forms of data transmission, IP telephony deals with a real-time application, which places additional demands on IT management. Issues such as prioritization of data packets and failover concepts play a decisive role, as does the measurement of all relevant characteristics pertaining to network and application behavior. This performance data forms the basis for planning, operating, and tuning the network as well as automating many IT administration tasks. Only an integrated suite like eHealth (developed by CA) can support the real-time demands of voice transmission without degrading the network service for data transmission. Such a suite also standardizes and simplifies the IP infrastructure in a given environment. Specifically, eHealth provides mechanisms that check and control all network components (hubs, routers, gateways, and switches, including inter-location traffic) and it expands IP networks to become bona fide multimedia transport networks.

Accordingly, when considering a conversion to IP telephony, one needs a migration strategy for the network and its management with the objective of standardizing IT management and simplifying it, despite the possible addition of new applications and requirements. This can only occur with the support of an integrated network management suite that meets real-time voice transmission requirements without deteriorating the network’s data transmission capability.

To that end, mechanisms are required to check and control all network components such as hubs, routers, gateways, and switches made by a slew of different manufacturers. These mechanisms then provide the VoIP traffic generated at the terminal devices with the required bandwidth in the corresponding quality so that all users can benefit from high-quality VoIP phone calls. In this way, existing IP networks can be transformed into multimedia transport networks because the entire VoIP traffic is provided with the required bandwidth.

The four-step migration

Executing this type of transformation successfully takes place in four phases. The process begins with an assessment of the existing IP network. Based upon an analysis of the network’s configuration for VoIP capability, one can then upgrade or replace the inadequate network components so as to have the required bandwidth and QoS functions from end to end.

With phases one and two out of the way, the third phase consists of implementing the VoIP solution, testing it, and putting it into operation step by step. Thanks to continuous monitoring, one can respond immediately to any quality deficiencies, by adjusting configurations, shifting loads within the network, or upgrading network components where bottlenecks occur.

Specific simulations of network behavior under maximum load conditions and a methodical analysis of the monitoring results accelerate VoIP implementation since the administrators are able to eliminate possible flaws from the onset before they result in failures. In addition, one can purchase solutions on the market that are able to determine from the simulation and monitoring results where the error sources are in the network or whether bandwidths require expansion.

After implementation, the fourth phase consists of proactive monitoring and management of the VoIP solution once it is in continuous operation. Monitoring and management tools are available on the market to verify and fine-tune IT performance as it relates to the VoIP system. These tools collect data that can be summarized a real-time status report regarding the status of voice transmission, or that can flow into a comprehensive IT management suite or are themselves part of an overall IT management solution. For example, a QoS module or a service-level management module can automatically ensure that the agreed upon service levels or quality characteristics are maintained system-wide for components as well as applications.

Embedded VoIP Solutions

Embedding VoIP solutions in a complete system and network management suite has one additional significant advantage over specialized programs. Specifically, the tools and modules of a complete suite can be used in all four phases of the VoIP migration, from evaluation all the way to operation, whereby the focal points of the application can be very different depending on the phase. Experts agree that the following three functions are especially critical to efficiently managing VoIP solutions:

1. End-to-end monitoring

2. Monitoring and management of QoS

3. Performance management of applications and services

Special VoIP features of an integrated system and network management solution might already come in handy when evaluating the IP network – a step that absolutely needs to precede the implementation of a VoIP solution. Compared to public telephone networks, current IP networks are still relatively unreliable and prone to performance fluctuations. In addition, compared to the line transmission of traditional telephony systems, packet-based transmission has its own unique error sources, which must be remedied from the start.

Assessments are essential

Assessments identify IP telephony-pertinent characteristics such as latency (transmission delays) and jitter (timing drift in the transmission of digital signals) as well as the utilization and availability of all on-hand IT components. The data network can be analyzed and evaluated with the aid of a performance management solution that measures the performance of all network, IT, and voice systems and generates profiles of the voice quality to be expected. Since voice quality is subjective, objective standards such as Mean Opinion Score (MOS) values are utilized to evaluate voice / image transmission quality.

In order to be meaningful, the assessment must look at the entire network handling a call. A single disruption or malfunction at one location can tangibly interfere with a phone call. It is critical that all components are matched to the intended use since the mentioned characteristics change under increasing loads – often abruptly.

The time and effort associated with performing an assessment cannot be simply avoided by making more than enough bandwidth available. That is a fallacy that may prove to be expensive and also pose risks. This approach may seem to work at the beginning, but it could also lead to unforeseen problems with increased use of the network and IP telephony because administrators lack sufficient knowledge about the network’s behavior.

These problems can be simply avoided by utilizing a network management solution for the assessment that is tailored or expanded to meet the special requirements of a VoIP application, and that also supports all subsequent implementation and fully operational phases of the VoIP migration. Thus, the IT team will be spared the need to buy and install new tools for the assessment and administrators will not have to repeatedly expend the effort to get acquainted with them.

End-to-end monitoring

A critical factor to the successful implementation of VoIP solutions is also to view the network as an end-to-end service. In order to ensure reliable voice transmission in day-to-day operations, proactive, real-time monitoring of all VoIP applications is essential. Otherwise determining the voice quality from the performance figures of the network components involved would be too time consuming and thus hardly a feasible alternative. For large-scale applications, the use of a “voice quality manager” for end-to-end monitoring and other tools for intelligent cause analysis and troubleshooting are helpful since they might pinpoint where the source of a problem might lie.

Such root cause analysis tools and end-to-end monitoring are useful for both test and planning purposes. For example, administrators can utilize a load generator to place a basic voice load on a network to then test a certain network route. Such load tests can later also be combined to enable administrators to run low-load, end-to-end tests on the entire voice network.

Based upon the measurement results, the tools determine characteristics such as the aforementioned mean opinion score. The measurement results may also lead to further simulation experiments, re-engineering of the network or voice transmission-related route changes.

Quality of Service monitoring and management
After the assessment and re-engineering have been executed, the convergent network is now capable of handling telephone calls in the specified quality. However, it must be consistently checked and managed, ideally by means of QoS mechanisms. An IP network is a dynamic, finite-bandwidth resource utilized by many applications and users. Voice and data transmission as well as other services compete for this limited resource, so it makes sense to have a higher level to handle load assignments.

QoS mechanisms can control various types of IP network traffic depending on their quantity and quality requirements – all within the context of a single, convergent network. It should be clear that supporting this critical function is essential in VoIP-related network components.

This process avoids problems caused by a dedicated allocation of resources for voice transmission. For example, if the resources for voice transmission fail, it would take considerable effort to redirect it to components reserved for data transmission.

It is certainly to one’s advantage when the QoS mechanism takes on the role of a dispatch service that can allocate all of the convergent network’s on-hand resources to the various types of traffic according to their service class. These service classes are defined by network administrators, provided with certain QoS metrics, and then controlled for each class according to the service level agreements (SLAs). As one gains more experience with network operations, this process will take place automatically as underlying policies are tailored to changes in service class or network loads.

Overlapping Performance Management of applications and services

This level of post-installation fine-tuning takes place on the basis of the VoIP service’s performance management reports. In this case, the traditional, component-oriented monitoring of network components and connections is no longer sufficient as already mentioned, since it does not allow problems pertaining to voice quality or service availability to be identified and remedied quickly enough. Rather, in-depth knowledge regarding the interaction of the numerous components is critical; real-time detection and alarm notification of bottlenecks or flaw as well as their preferably automatic correction are also essential. More added value is offered by proactive flaw-detection solutions that thereby substantially increase availability and service quality.

Overall, the standardized and simplified management of the entire IT infrastructure helps administrators to avoid excessive costs and unnecessary compromises in generating good voice quality. To make this happen requires the integrated interaction of diverse management skills, seamless end-to-end monitoring, specific checking of quality of service, and comprehensive performance management of applications and services. When combined with intelligent root cause analysis as well as the required test and simulation tools, planning and implementing IP telephony in a company should be straightforward and problem-free.

Author: Juergen Magiera, CA.

Additional Information

Survey results

Managing 21st century networks – a world of convergence
A research carried out by Quocirca on behalf of CA shows that Europe’s IT leaders are concerned that they do not have the resources to proactively manage key network technologies such as MPLS (Multi-Protocol Label Switching) and VoIP (Voice-over-IP), and lack quick access to information on problems, their causes, who’s impacted and how to fix them.

See full report here:
http://ca.com/de/press/converged_network_management.pdf

White Papers
• Managing QoS Policies Means More Business
This white paper details how you can harness CA's detailed reporting and analysis of QoS to create a competitive advantage in deploying QoS to market while creating ROI.

See Full White Paper here:
http://www3.ca.com/Files/WhitePapers/qos_wp.pdf

• Managing Voice Over IP for Successful Convergence
Voice over IP (VoIP) has much to offer in cost savings but some customers have concerns about VoIP call quality compared to the quality of traditional voice services. This white paper will help you learn how to take the right steps so that voice quality is assured. It leads you through the pre-deployment assessment and describes how quality of service should be applied to ensure your production network will meet users' expectations.

See Full White Paper here:
http://www3.ca.com/Files/WhitePapers/manage_voip_success_converg_wp_en_us.pdf

Wikipedia Links
VoIP: http://en.wikipedia.org/wiki/Voip
Quality of service: http://en.wikipedia.org/wiki/Qos
Network performance management: http://en.wikipedia.org/wiki/Network_performance_management
Mean opinion score (MOS): http://en.wikipedia.org/wiki/Mean_Opinion_Score