Problem solve Get help with specific problems with your technologies, process and projects.

Hosted VoIP: Take the headache and heartache out of VoIP, part 2

No matter what size network an enterprise is running these days, large and small alike are faced with increasing technical and financial challenges. The driving force behind this shift is new market innovations, such as the convergence of voice and data, which makes network operations more complex

Hosted VoIP: Take the headache and heartache out of VoIP, part 1

The good news is that VoIP has been around for a number of years, so some of its major risks are well understood. The industry is constantly developing best practices and tools and techniques for protecting and controlling VoIP networks. For instance, the International Organization for Standardization offers ISO 17799, which provides recommendations for information security management, also provides a common basis for developing organizational security standards. Similarly, the International Telecommunications Union's X.805 standard, pioneered by Bell Labs, defines security architecture for systems providing end-to-end communications. And NRIC, the Network Reliability and Interoperability Council, provides best practices guidance in a number of areas that relate to VoIP operations.

While the tools to deal with security concerns and VoIP exist, the time it takes to implement them may be the harder part of the equation. Because the industry is constantly evolving, security issues may pull away hours and hours of resource time from already strapped IT staffs.

Even the creation of a best practices model for an individual company is a laborious process. Typically these includes both and internal review that involved detailed -- users, administrators, managers and other employees -- as well as the more traditional security equipment assessments to show potential technical security gaps. A well-devised security plan must be able to include these unforeseeable risks and minimise them – however finding the time can be the trickiest part. In a hosted environment much of this work, especially keeping current with new types of attacks and solutions, will already be addressed.

The second major issue that is also commonly overlooked or minimized in the zealous race to move to a VoIP network is QoS. While QoS has long been discussed in data circles, when an IP network is carrying voice traffic it becomes an even more critical element. Just as planning is a key aspect to controlling the potential security issues; planning and network design are the foundation to building a VoIP network that delivers upon quality expectations. A good network design methodology includes prioritisation, traffic engineering, a plan to handle voice in a converged network and a restoration process.

In addition administrators must put a call admission control (CAC) process in place. CAC allows the customers to properly design the network to carry the traffic load even at the busiest times while still meeting QoS objectives. This complex process can be managed, for the most part by using one of the following methods: per call bandwidth reservation, local measurement based management, path-based management and link-based management. A network needs to carefully consider each choice, because the wrong one can have a direct impact on total voice quality.

In a traditional circuit based network voice is given a dedicated bandwidth allotment so the quality is assured. In a packet environment a customer must either predict voice quality in a new VoIP deployment or assess voice quality in an existing network to ensure end user satisfaction. Both scenarios can be managed by using a well-designed modelling process. One way to accomplish this is to combine both subjective testing with objective testing models. The subjective category includes the e-model scale of user satisfaction categories (ITU G.109) that set definitive end values and give a Mean Opinion Score (MOS). The objective components then measures network impairments such as delays, packet loss and echo, and computes a total score. In the case of a brand-new network these models are then combined with a network performance prediction tool (NPPT) that takes into account network information, VoIP traffic demand patterns and a network performance prediction algorithm to deliver a VoIP voice quality prediction report.

In the case of an existing network the e-model testing is still used for a voice auality assessment portion, but it is combined with other assessment tools such as network discovery, network performance measurement, delay asessment plan, delay assessment and root-cause analysis to deliver a VoIP voice quality assessment report.

While the industry has a number of different approaches to QoS it is clear that voice quality is key to the ultimate success of a VoIP deployment. The planning and traffic prediction make or break the deployment. The use of a hosted environment where detailed planning was built into the network design and will continue to be upgraded and monitored as new technology emerges may make a significant difference in overall performance.

Given today's competitive operating environments, the benefits of the hosted model provide some compelling reasons to consider it as a viable option. These may include lower overall operating expenses, ability to provide enhanced security, improved network performance and ability to quickly rollout new services such as VoIP. This approach allows enterprises to focus on their own core network competencies and re-deploy staff to areas that make the most use of their expertise.

As VoIP and a wide variety of other broadband communications applications including unified messaging, security and mobile extension applications that can extend the functionality of the office phone system to a mobile environment, continue to emerge on the scene, the network landscape is dramatically changing. The continuous push to add new technology is tipping the scales in what was once a closely guarded possession for organizations -- network operation. When customers combine the cost-effectiveness and the ability to immediately offer new services the decision to outsource is becoming clearer every day to both the network operator and enterprise CIO.

About the author
Andy Williams is President of Lucent Technologies, Europe. He joined Lucent in August 2005 after a 25-year career with IBM, where he held a number of senior leadership positions, most recently that of general manager public sector EMEA within IBM Global Services.

As President of Lucent Europe, Williams is the senior executive for the company based in Europe. He is directly responsible for sales activity in Europe and will be leading the company's growth plans for the region, powered by innovation from its research and development arm, Bell Labs. Lucent employs around 4,700 specialists in Europe, serving most of the major service providers and telecommunications companies.

Williams holds degrees in Mathematics from Cambridge University, U.K., and is married with two children. He is based in London, but travels extensively across the region, supporting Lucent's operations at 44 locations in 14 countries.

Dig Deeper on VoIP Migration and Implementation

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.