The Dreadnoughts: Enterprise Class PBXs

As of this writing, the majority of organizations with thousands of employees at one location continue to use traditional, TDM enterprise class PBXs. For the next few years, these mainframe-like machines will continue to dominate the seas of telephony. They have extremely high reliability, more features than any person with a real job would ever use, and a vast flotilla of accessories/applications. The large installations cost millions.

Business needs should dominate technology "coolness." If the organization must have a new telephone system with all the advantages of mature technology, then cost analysis shifts away from technical architecture and toward features, needs, and smart planning. In evaluating competing vendors such as Avaya, Siemens, Nortel, and others that play in the large PBX world, the following should be included in the analysis:

• Has an accurate count of users, stations, and non-voice devices such as modems and faxes been obtained? Organizations often overbuy because the actual number of necessary lines is not known. Analog lines (for faxes and modems) are particularly prone to excess.

• Do voicemail and other proprietary links match the organization's infrastructure? For example, will the PBX vendor's unified messaging offering work with the installed e-mail system?

• What functions are really required for each station? Some organizations buy low-end telephones for lower-level employees and then upgrade as they are promoted. Sometimes, the cost of this transition can exceed the cost of providing a standard, strong feature set telephone for all employees, with only individuals in specialty positions having "more buttons."

• How much expansion room is available? Traditional PBXs scale well but usually in steps rather than on a continuous curve of users versus costs. Once a shelf is filled up, for example, a new shelf and cards must be purchased, at a substantial incremental price. Of course, most of the big PBXs network with one another easily (as long as they are from the same manufacturer).

• How much of the implementation effort is the vendor willing to do? Implementation is a significant human and financial drain, so responsibilities should be drawn early. Assume for example that the firm is converting from Siemens to Avaya technology. Understanding what each user needs to have programmed into the station and how that relates to the new system (station programming) is a big job. The translation is not straight-forward and requires considerable manual effort.

• Do the PBX and related systems support hoteling — a setup that allows employees infrequently in the office to direct the system (usually via a kiosk) to ring their assigned number at the telephone they are using that day?

Having discussed premise equipment or alternatives, we can turn to public network access and related cost reduction.

Small Offices and SOHO Markets: Key Systems versus PBX

For many small offices, a PBX is unnecessarily large and complex. Key systems will serve offices with less than 60 telephones; but beyond those numbers, a PBX is required. From a financial perspective, the "key" decision (pun unintended) is whether an office or plant will grow significantly, thus requiring a forklift upgrade to a PBX. Key systems cannot be easily scaled, although the distinctions between PBXs and key systems are less distinct now than in the past. Following are the traditional points of separation between key systems and PBXs:

• Each key system telephone has buttons or keys that are used to directly access external lines ("Harry, pick up line 24"). In a PBX environment, users are not linked to specific external lines, although they may have a second or third "line" on their telephone set. A key system may have 32 telephones and a dozen external phone lines.

• PBXs share trunking lines among a larger number of users. Using a DID (direct inward dial) number, a call is routed from an incoming trunk to the user's telephone. The PBX has the ability to switch rather than depend on a one-to-one user:line connection. This allows a much higher ratio of users to trunks and, with appropriate equipment, permits scaling into the tens of thousands of stations.

• Although key systems continue to offer more features, PBXs still provide considerably more functionality, both from an end user and internal perspective.

• On a per-station basis, key systems can cost as little as $200 to $250; that number easily doubles for PBXs.

• The ability to select trunks can reduce costs in some areas of the country where LECs charge higher rates for pooled access trunks (i.e., those used by PBXs) than for lines that are selected by the user.

• PBXs have a robust ARS (automatic route selection) capability. By looking at part or all of the numbers dialed, the PBX can select the most cost-efficient method of completing the call. For example, assume that a company has a headquarters building in Chicago and a smaller office in Denver. A user in Chicago dials the full ten digits to reach a co-worker in Denver. By looking at all ten digits, the PBX determines that the call should be routed via tie lines (or the IP equivalent if VoIP is in place) rather than the public network. It is important to have a PBX that can look at all ten digits because just looking at the area code and exchange may not be sufficient. For example, the Denver office might only have part of the exchange — the rest belongs to other organizations in the same area code.

• PBXs generally support full networking, including voicemail and a uniform dial plan (a uniform dial plan allows, for example, five-digit dialing across the United States and even worldwide; employees can reach each other with a minimum of dialing).

Key systems increasingly offer robust features that make them more attractive to the smaller office. In addition to their low cost, they offer functions such as:

• Voicemail

• Hunt groups. Incoming calls can be routed to a workgroup. Individuals will receive the calls in a predefined order and the call will be sent to the next available person in the sequence. In these cases, employee 1 on the hunt group list works hardest; employee 5 works least. Some key systems use smarter algorithms to distribute calls more evenly.

• Auto-attendant. Callers are sent to a main receptionist who then distributes the call as required. Computer-based auto-attendants provide the ubiquitous "Press 1 for Sally, 2 for Fred, etc."

The high end of the key system range is now the target of IP telephony vendors. However, the low end, say three to ten stations, continues to be dominated by traditional key system vendors. Because the low-end IP PBXs still require a reasonably robust server, there is a fixed cost that is difficult to overcome at the small end of the spectrum.

An example of the inexpensive (but not low-quality) key system is the Cortelco Aries 308 (see Exhibit 1). It accepts three CO (central office) lines and serves eight digital stations. Options for paging, external music on hold, and remote programming of features are available. If a small office needs only basic services, no inter-office system linking, and there are no rapid expansion plans, such a system may be appropriate.

Exhibit 1: Cortelco Aries 308 Key System

Some key systems include broadband connections. For example, BizFon's 680 KSU (key system unit) has a 16-channel DSL card. Others support standard BRI (basic rate interface) and PRI (primary rate interface) links that allow more channels and potentially a discounted long-distance rate. The discounted rate is due to the ability of PRI trunks to directly connect to the long distance carrier, bypassing the local telephone company (this is sometimes termed "dedicated" access).