Network bandwidth and usage are only two of many issues to consider before implementing an IP telephony solution.

While it's true that toll-call charges, both nationally and internationally, have come down in the past few years, most new phone systems being installed today are not traditional TDM-based, circuit-switched PBX systems, but instead IP voice and data networks.

The most common anticipated benefits of VOIP implementations include lower total operating costs, enhanced end-user productivity, improved IT organization efficiency, and reinforced market differentiation and brand image. To achieve these results, an organization needs to move toward unified communications, and an IP telephony service is a first step. As an IT manager, the question is this: are you ready--and is your network ready--to flawlessly move into the future era of converged communications? Or will you be the one holding back your organization because you fear the complexities of layering voice communications onto your data networks?

Once you deploy VOIP in your data network, there's no doubt that it becomes the most critical application on the network. Most users won't care much if their email server is down for 30 minutes, but the CEO and the rest of the employees aren't going to tolerate dropped in-process calls or absence of dial tone. If there is one thing we have come to rely upon today in the era of technology, it's a reliable phone system. Take that away from someone, and you're creating a hostile group that won't let you sleep at night.

Therefore, implementing a VOIP system is a balancing act between risks and rewards. Given that businesses today simply must reduce costs, there is going to be a greater tendency to take calculated risks in order to achieve reduced costs. One major contribution that an IT department can make to an organization is overseeing the smooth implementation of a VOIP system in place of a traditional and redundant TDM system. The challenge is doing this while maintaining system reliability and availability, equivalent voice quality to the TDM system, easy scalability, multivendor interoperability, a full suite of communications features, and a favorable overall cost and payback interval. That may sound like a tall order, but the technology solutions of today make it possible. To reach this goal, detailed planning, good vendor support, and a steady hand are all that's required, and the results make it worthwhile.

The initial question in the mind of every IT manager considering adoption of a VOIP solution is whether the existing infrastructure can handle it. The bottom line is that most companies upgrade their networks before they roll out a VOIP system by adding smart switches that can route traffic intelligently or software that can manage and prioritize voice packets over data. They will consider dual-core switching and premium VLANs to segment traffic and allow for uninterrupted uptime, will install redundant power supplies, and will do whatever it takes to ensure servers running the IP phone system stay up during power outages. Server reliability is a key factor in an IP telephony system, so reliable IBM Power Systems are a natural choice for running an IP application. Many organizations will have additional capacity on their server that can be re-purposed to this end by running either the 3Com or Nortel solutions for VOIP on the IBM i.

There will always be the question of whether you want to have a homogenous network to avoid finger-pointing when the phone system underperforms. Cisco has capitalized on this concern. Since most networks already use Cisco routers and other hardware to run the data network, the company has invested heavily in VOIP technology to offer users the assurance of a single-source supplier. The fact is, however, that the single-vendor pitch is largely hype. A Cisco network can carry and prioritize IPT traffic from any product that is based on industry standards. What is probably more important than uniform sourced equipment is a reliable and highly skilled VAR and system integrator. Choose the wrong integrator, and you will wish you had stayed with your old TDM system.

Enterprises often focus on the product and its features. However, feature sets are remarkably similar today in terms of what they do and how they function. Surely, one should ask whether a homogenous IPT and data network is necessary, but the focus on long-term viability of the vendor in the IPT space is more important than feature sets. Focusing too closely on the product line can easily result in a flawed implementation. Strength of the vendor and the experience and expertise of the integrator are generally more important. The choice of an integrator can make or break an IPT deployment.

When choosing a vendor, look for one that has made a significant commitment to IPT research and development and has demonstrated market leadership. Obviously, purchasing a product from a vendor that is a candidate for a buyout or is financially unstable could cause major problems down the road. Things to consider when choosing an integrator might include the number of IPT implementations they have completed, the types of environments they have worked in, whether they have references from similar implementations, whether they offer an on-site visit to a similar deployment, and whether they will provide certifications for all integration staff doing the work.

Many variables impact the quality of an IP telephony system: the IP-PBX solution used for voice communication, the network equipment to carry voice and data traffic, the amount of bandwidth available to all users at all sites, the total amount of traffic moving across the network--all these will affect performance. Generally, voice quality is not an issue over a corporate LAN because there's usually plenty of bandwidth. The congestion point comes at the boundary between the corporate LAN and the WAN. It is here that voice and data traffic must be carefully controlled, as it's tantamount to freeway traffic feeding into a two-lane secondary road. If traffic isn't controlled, performance issues will result. These likely will emerge in three critical areas. The first is latency, which is the delay in delivering the voice stream from the speaker to the listener. The second area is jitter, or the variable delay in the delivery of each voice packet. The third area is packet loss and dropping of individual packets caused by network congestion. Each of these three issues can cause degradation of voice quality and compromised system reliability.

VOlP is real-time two-way communications and is sensitive to network delays. VOIP requires a latency of no more than 80ms each way for good voice quality. Delays of 150-180ms each way still might produce acceptable voice quality, depending on the protocols employed. Network devices attempt to address latency with various queuing techniques to ensure that voice packets take priority over other traffic. That may not be sufficient, however, and more-stringent bandwidth management--allocating a specific amount of bandwidth to each type of traffic--may be required.

Jitter causes irregularities in the delivery and flow of data and can wreak havoc with real-time applications like VOIP. If jitter causes delays beyond 20 to 30ms, there will be a degradation in voice quality. Routers offer queuing mechanisms to smooth out the delivery of voice packets, and some vendors have their own jitter buffers. Packet loss can be a serious problem at the LAN and WAN boundary, where there is generally contention for bandwidth. A packet loss of 1 percent or less in general will not result in a degradation of voice quality, but when it reaches 3 percent or greater, the listener will notice the conversation breaking up, and the call might even be dropped. Traffic control and performance optimization solutions at the LAN/WAN boundary, utilization of bandwidth management technologies to minimize IP congestion, and the maximizing of application performance over the WAN are all ways to address these problems. Needless to say, having visibility of all WAN application traffic is essential to understanding what's happening.

Having good tools that deliver granular diagnostics and reporting capabilities will allow a network administrator or vendor to troubleshoot potential VOIP issues. Software-based tools may offer advantages because they allow faster network assessment than hardware-based tools, which require plugging them into different points of the physical network. Software may also support advanced simulation capabilities, allowing an administrator to test actual data application traffic in order to determine the maximum acceptable quality VOIP capacity that an enterprise can handle in the presence of real application traffic. If relying on a vendor, ask what kinds of tools they're using and what parameters they will use to test the network before deploying the IP solution.

Before deploying VOIP, IT managers must take into account the impact of voice on other applications that are running on the network. Simulating high-traffic periods, such as end-of-month reporting, is important to a reliable VOIP implementation. Since enterprise networks are always changing, regular reassessment on a quarterly or monthly basis should reveal trends that can prove vital in predicting problems. Scheduling the VOIP test system to generate and measure a short call every hour between locations can help to identify outages and reduced voice quality. Administrators may wish to implement continuous real-time monitoring that can inspect the actual IP call traffic down to every single IP call that occurs.

While there are risks in implementing a VOIP solution, they are no greater than the risks in implementing any other mission-critical application on the network. IT managers can take heart in the knowledge that more than 85 percent of system administrators who were surveyed recently reported their VOIP implementations met or exceeded their expectations. The objective, then, is to avoid being lumped among the 15 percent who reported they were not happy with the outcome. Through proper planning and the support of qualified vendors and integrators, chances are good you will be among that happy, satisfied 85 percent.

Chris Smith was the Senior News Editor at MC Press Online from 2007 to 2012 and was responsible for the news content on the company's Web site. Chris has been writing about the IBM midrange industry since 1992 when he signed on with Duke Communications as West Coast Editor of News 3X/400. With a bachelor's from the University of California at Berkeley, where he majored in English and minored in Journalism, and a master's in Journalism from the University of Colorado, Boulder, Chris later studied computer programming and AS/400 operations at Long Beach City College. An award-winning writer with two Maggie Awards, four business books, and a collection of poetry to his credit, Chris began his newspaper career as a reporter in northern California, later worked as night city editor for the Rocky Mountain News in Denver, and went on to edit a national cable television trade magazine. He was Communications Manager for McDonnell Douglas Corp. in Long Beach, Calif., before it merged with Boeing, and oversaw implementation of the company's first IBM desktop publishing system there. An editor for MC Press Online since 2007, Chris has authored some 300 articles on a broad range of topics surrounding the IBM midrange platform that have appeared in the company's eight industry-leading newsletters. He can be reached at This email address is being protected from spambots. You need JavaScript enabled to view it..

Also Read

Practical Middleware: Architecting for the Future

Practical Middleware: Starting the Journey

IBM i 7.2: The Carrot Versus the Stick