August 2001

BY DICK HAYTER

To determine where services should "live" in the network, several factors must be considered. As the number of subscribers increases, the overall complexity and need for capacity increases. The ability to maintain and manage applications becomes critical as the number of users and services grow. And, finally, the applications must be able to interoperate fully with existing legacy networks and users.

The reality is that there will probably be no single point of service deployment in next-generation networks. There are benefits to be realized from all of the options, and it's likely that service deployment will be a hybrid of the three. It's going to take a "divide and conquer" approach to replicate the services that have taken several decades to develop in the circuit-world.

At the Edge
Devices at the network's edge require two components to deliver services to the end-user -- a media handler and "smart" software to handle services. Packet-based devices like SIP phones, designed as "smart" end points, combine those essential components into a single device. Black phones, which just handle media, can be "educated" by giving them access to intelligence through customer premise equipment like PBXs or desktop PCs.

It's logical to deploy simple services such as two- or three-way conference calls from smart end points. The less-complex applications do not drain network resources and are easy for customers to manage from their desktop.

When to Take the Next Step
As the service becomes more complex, the number of services grows, or the number of users increases, it's time to consider moving the service deployment site further into the network. While it's relatively easy for a smart end point to handle a conference call with two or three legs, the situation changes dramatically when a user wants to conference in 10 or 20 people. If a service is deployed from a large number of end points, it becomes increasingly complex to maintain and update. As the level of complexity increases, end customers are less inclined to manage the feature from their desktop. And, service management for the provider becomes more cumbersome.

To avoid a drain on network resources and simplify management, it makes sense to move complex applications into the network. This approach allows providers to take advantage of economies of scale by sharing the application or service among many subscribers.

The Network Options
Once the decision is made to locate a service within the network, several issues have to be addressed. Will the application reside on a softswitch or an application server? And, who will control the service delivery -- carriers or third-party application service providers (ASPs)?

The softswitch's role in the new network architecture makes it a logical point from which carriers deploy services when the service needs to interact with the PSTN. It serves as the mediator between the PSTN and the IP networks providing access to voice and data end points. It supports both intelligent network and SIP call models. The SIP User Agent (UA) hosted on the switch provides "intelligence" to black phones in the PSTN and emulates SIP behavior on their behalf. Extending the SIP UA to include value-added service becomes a natural fit for the softswitch.

Services, which require specialized resources and service logic, are likely to find a home on dedicated application servers. Enhanced media processing such as multi-party conferencing and text-to-speech conversion can be provided using specialized media servers under the control of the application servers. This approach allows providers to customize and quickly deploy services to match the unique needs of particular customer segments.

ASPs offer a new model for service delivery and bring innovation to the market. Third-party development encourages open, non-proprietary applications and opens new distribution channels to businesses and end users. With the ability to purchase services from many providers, users can get the best application fit by choosing exactly the services they need to meet their unique requirements.

No matter what architecture prevails, there are common challenges. Feature interaction has been a major problem in the PSTN. With more features, multiple players, and multiple media forms, one only expects this problem to get more complicated. A modular, scalable signaling network coupled with service creation and multiple distribution systems is the foundation required before the dream of innovation can be fully realized.

Richard (Dick) Hayter is the assistant vice president of marketing for the Network Systems Division of Tekelec. Tekelec, a leading supplier of signaling and controls systems, develops innovative network switching and diagnostic solutions enabling the convergence of traditional and converged wireline, wireless, and IP voice and data communications networks. The company also provides products and solutions for call centers and other telecommunications markets. 

[ Return To The August 2001 Table Of Contents ]