This chapter assumes you are adding VoIP to a reasonably mature data network, and that you are already managing your network components, your users, and their applications. Deploying VoIP means you are adding a new, complex application to your network. It is not the intent of this chapter to reteach network and application management; instead, it discusses how your management goals and tasks will change after VoIP is running on the network.
In the deployment stages of your VoIP project, your goal was to get it right the first time. You needed to make your first group of VoIP phone users happy, with high levels of availability and call quality. But now, in the management stages, your goal is to keep those users happy by keeping availability and call quality acceptably high. What is more, you will be doing this in the face of ongoing changes: more phone users, new network applications, different network components, and daily security intrusions.
Keeping your VoIP system healthy involves carefully managing four things well:
Operations Smoothly handling the day-to-day changes to the network, applications, and users
Availability Ensuring high uptime for the overall phone system
Call quality Making sure that every phone call sounds good
Accounting Ensuring that the calls are being charged properly, to the right people and departments
Some aspect of security is integral to each of these four topics, so it has not been broken out into its own category. For example, security intimately affects operations. You want tight control over who is permitted to make changes to any VoIP components. Unauthorized access to the phone system or data network must be prevented because it can affect availability and quality through degraded performance. Security also affects accounting. Telephone records contain private informationwho called whom, for how long, and when. In addition, Chapter 8, "VoIP Security," is, as the title suggests, completely devoted to VoIP security.
VoIP management encompasses a broad set of tools, techniques, and processes to ensure the reliability and availability of a VoIP implementation. With all the effort required to get VoIP up and running, it is easy for management to be forgotten or postponed. However, good VoIP management practices are not optional in a successful VoIP deployment. There is an ongoing cost associated with VoIP management, so it helps to understand why it is so important, as explained in the following section.
Why VoIP Management Is Essential
VoIP will be a business-critical application on your networkan application that your business depends on for day-to-day operations. Although it is true that some businesses depend on the phone more than others, when you pick up the phone, you expect to hear a dial tone every time. When you make a call, you expect good qualityalways. VoIP management is required to keep a successful VoIP deployment successful.
Deploying VoIP can be a challenge, but after it is up and running, the challenge is not over. A good management system can simplify the day-to-day operations of a VoIP implementation. Consider some of the benefits of VoIP management:
Toll-quality telephone calls How do you know if the call quality is good? Most phone users will promptly let you know if the quality is poor. The distinction between good and bad is readily measured, using the mean opinion score (MOS) as a standard. A good management system can help you recognize performance trends that can lead to lower-quality calls. Management tools can describe precisely what the current quality isgood, acceptable, or poor.
High expectations for reliability Through years of Public Switched Telephone Network (PSTN) usage, people have become accustomed to a high degree of availability from their phone system. In fact, most carriers provide approximately five nines of availabilitythat is, a dial tone is available 99.999 percent of the time when you pick up a handset. How can you meet these expectations? You need a management system in place to maintain high availability.
Avoiding failures altogether The best way to achieve low downtime is to avoid failures proactively. Deploy hardware correctly, watch trends, replace failing components before they cause a crash, and upgrade to eliminate bottlenecks before they degrade quality (such as CPUs, bandwidth, routers). It is better to avoid problems entirely than to be very good at finding and fixing themalthough good management tools can help you do both.
Knowing when failures occur With so many components in your VoIP system, when a component fails, it is doubly important to know whether it is a gateway that provides access to the PSTN or a critical server that routes all of your VoIP calls. It is impractical to manually inspect all components for failures. A good management system monitors and detects failures in key components. Once a failure is detected, management tools also can take automated actions, such as notifying IT staff of the problem via pager or e-mail.
Pinpointing and diagnosing problems quickly Once you know that a failure has occurred, the next step is to locate and then diagnose the problem. Diagnosis can be difficult; you may know that no new phone calls can be made, but you may not know that it is happening because a software application has stopped on a VoIP server. A good management system can help provide quick problem resolution to ensure customer satisfaction.
Accomplishing moves, adds, and changes (MACs) smoothly They need to be done in a timely manner that is nondisruptive to other users and other network applications, and they must not compromise existing security measures.
Maintaining privacy In the U.S., it normally takes a search warrant to examine the telephone records of an individual or an enterprise. Who was called, by whom, and when the call was made are all considered private information. In a VoIP system, call detail records (CDRs) containing this information may sit in the database of a VoIP server on your premises. They need to be protected from unauthorized accessonly approved individuals or programs should be able to read, delete, or modify this data.
Filtering relevant management data VoIP servers and devices can generate large amounts of data in log files and CDRs. This information can be very useful, but the sheer volume of data can make it difficult to find what you are looking for. A good management system can quickly filter large amounts of performance data to find the relevant information.
Ensuring that regularly scheduled system housekeeping tasks are run consistently VoIP systems comprise web servers and database systems, which typically have jobs automatically scheduled to perform backups, data archiving, and database grooming. You don't want to be surprised to find that your backup, at the moment you need it, has failed for the past six months.
Planning for future upgrades and purchases What is your budget and what is your purchase schedule as the network, users, and applications change over time?
This list focused on VoIP's technical challenges, which is where your IT team is also most focused. But there is probably a management team in your organization that has a focus not on the details of the technology, but rather on the bottom line.
A Look at Downtime Costs
The previous section introduced many areas where VoIP management offers obvious benefits. This section picks one of these areas, availability, and looks at the costs of managing it poorly.
When the phone system is down, the associated costs from lost revenue, lost productivity, and lost time for your IT staff immediately begin to mount. A 2002 TeleChoice study found that the lost revenue per employee per hour can reach more than US$40 (for industries heavily dependent on phone usage) when the phone quality is poor or the system is unavailable. A quick calculation shows that the costs can quickly add up as the number of employees increases. Take, for example, a company with 5000 employees and 5 percent average downtime (over a one-week period):
5000 employees * $40/hour/employee * 0.05 = $10,000 lost revenue/hour
Given an 8-hour workday and 40-hour workweek, this would translate into a cost of US$400,000 for the week. If phone quality is poor or unavailable for a greater percentage than this, the cost is even greater.
A Find/SVP survey conducted in 2000 found that the average network outage among Fortune 1000 companies lasted four hours and cost US$330,000. The same survey showed that "a typical company experienced nine outages per year, resulting in annual losses of almost US$3 million (excluding the cost of lost employee productivity)"and possibly excluding the cost of lost sales due to unavailable telephone service.
Another cost of downtime is borne by your IT staff. The time it takes to constantly fight fires and deal with problems can take its toll on productivity. You want your staff to be able to focus on making enhancements to your infrastructure rather than constantly putting out fires. A good VoIP management system can pay for itself in the time it saves staff and the productivity enhancements it can provide. The goal is to provide a self-healing systemone that can automatically fix as many problems as possible versus having an IT staff person to resolve every problem.
A similar cost discussion can be developed for other areas that will suffer as a result of poor VoIP management: call quality, privacy, MACs, other network applications, and so on. Rather than creating an extended list of cost justifications for good VoIP management, the next section moves on to one more clear challenge: the inevitable growth in usage that will occur when your VoIP deployment is successful.
No doubt about it, a VoIP system will change after it is initially deployed. New users may be added and existing users may change locations. Easy moves, adds, and changes are among the key benefits that VoIP can provide over a traditional PBX. But MACs require that procedures be carefully followed. Provisioning needs to be done for each new user: new users need a telephone number and IP address; their calls will generate traffic that requires network bandwidth; their call records will occupy database disk space; and so on. VoIP management helps with provisioning for new users.
In addition, a VoIP management system can let you know if specific servers, gateways, or links reach their capacity. If, for example, you have a WAN link that is fully utilized, you need to delay adding downstream users who may increase call traffic until you have bumped up the available bandwidth.
VoIP systems can really be taxing for IT managers. More users and more network applications continually eat up the consumables: bandwidth, addresses, ports, CPU utilization, RAM, and disk space. As an IT manager, you add more of these consumables to the right places incrementally, as necessary to make system performance acceptable. For example, you may have enough RAM in a router when a VoIP system is first deployed, but as traffic increases over time, call quality may decline unacceptably. If you then double the RAM, performance improves, and you have bought yourself some more time for the traffic volume to grow.
VoIP management also can help establish trends showing network behavior and performance over time, so you can tune your existing infrastructure and plan your spending. As you expand or change your existing system, you return to the top of the IT life cycle chart again, doing planning and analysis for the improvements.
So far you have been provided with a sense of why VoIP management is an integral part of your VoIP project. Next you take a look at what needs to be managed. If you examine the broad range of components that need to interact correctly for VoIP to function, you will see many potential points of failure. The next section addresses which components need to be managed and monitored. The core components of a comprehensive VoIP management infrastructure are introduced, starting with a discussion of operations management.