Enterprise customers and medium-sized and large businesses enjoy the benefits of a highly competitive market for telecom services today. The volume of traffic generated by these customers justifies the use of dedicated circuits from customer premises to multiple local and long-distance networks; and the PBXs that are used by these customers provide call routing intelligence that directs traffic on a call-by-call basis to the most appropriate carrier. The telephony services purchased by these customers typically relate to the basic transport of voice between multiple enterprise locations and the connection of voice traffic into various public networks.

The small business or upscale residential customer with fewer than 16 lines does not have the advantage of multiple competitive service offerings. The economics of access do not support direct connections from these kinds of customers into long-distance networks. They are, therefore, dependent on a single incumbent service provider for local dial tone and for access to all other services.

These market segments generate the bulk of profits for incumbent local service providers. Service providers earn large amounts of revenue from per-minute charges on local calls, from origination and termination of long-distance calls, and from usage of custom calling features and other local services, such as voice mail. And because there has been no effective competition in these market segments, tariffs are high relative to costs, and margins are very healthy.

Today, all local voice service is delivered from traditional circuit-based, local-exchange switches, largely because no other solution exists. This represents a major barrier to local voice services for the following three key reasons.

Cost of Local-Exchange Switching Solutions

The local-exchange switches market is dominated right now by a few powerful players who have built highly profitable businesses serving the needs of incumbent local service providers. Local-exchange switches from these vendors are optimized to suit end offices that serve many tens of thousands of lines. While the scalability of these devices is not in doubt, high common equipment costs make these solutions wholly unsuited for smaller deployments where only a few thousand lines are needed. An entry-level local-exchange switch typically costs on the order of $5 million, an amount that seems certain to deter competitive service providers from entering any but the largest markets.

Some competitive service providers have succeeded in entering second- and third-tier markets with traditional local-exchange switches by installing a single switch to serve multiple cities. The switch is logically subdivided into multiple local-exchange codes, one or more for each city, and connected by trunking facilities to local access networks in each city. While this approach does deliver lower switch costs per line served, any savings realized in switching hardware must be offset against the very substantial additional transmission costs of backhauling traffic from each city to a central switch location, and then sending all local call traffic back to the city from which it originated.

If local-exchange switching solutions were available at lower entry-level costs than traditional circuit switches, competition in local voice service would be stimulated, and customers would benefit from a wider choice of service providers and lower tariffs.

Lack of Service Differentiation

All the traditional local-exchange switches offer the same broad set of features for custom calling services: call waiting, call forwarding, caller I.D., selective call blocking, etc. Most of these features have been available for a number of years, and the delivery of completely new features is now relatively rare. This is partly because it is so expensive to develop and test new features and partly because the feature set available today comprises virtually every feature it is possible to imagine a subscriber being able to program from a telephone keypad.

Since both incumbents and competitive local service providers are dependent today on the same limited set of local-exchange switch products, they are constrained to offering exactly the same set of services. In these circumstances, the only way for a competitive service provider to win customers is to offer lower rates.

Differentiation solely on price has not proven to be a good long-term business strategy in the telecom business. If local-exchange switching solutions were available on which genuinely new and attractive features could be offered, then service providers would have a chance to differentiate on services rather than on price. This could offer a much brighter outlook both for margins and for customer retention.

Barrier to Network Migration

Traditional local-exchange switches are all based on circuit switching techniques. Within the switch fabric, voice traffic is represented as 64kbps streams. At the input and output ports of the switch, the 64kbps streams are time division multiplexed into higher-speed digital facilities. The intelligence of the switch that performs call routing and feature processing is integrated tightly with the circuit-switching fabric.

The economic advantages of packet voice are driving both the access and core voice networks away from circuit switching and toward packet. As packet voice becomes widely adopted for both access and core networking, the traditional local-exchange switch represents an island of circuit switching that connects these two packet voice networks. The packet-to-circuit conversion that must be carried out at both input and output of the local-exchange switch, however, introduces undesirable additional cost and transmission delays into the voice path.

If a local-exchange switching solution were available that were capable of delivering local voice services and custom calling features directly over a packet-switching infrastructure, then unnecessary packet-to-circuit conversions could be avoided. This has the dual effects of reducing cost and improving quality, and it moves the voice network a major step closer to the ultimate goal—homogeneous end-to-end packet voice.