9.1.1 Distributed Intelligence

In the past, the intelligence within networks needed to set up and release calls was typically distributed across the local and transit nodes. Each node had routing tables and information on supplementary services to facilitate the control of calls. This architecture provides a secure foundation for communications, but it suffers from a lack of flexibility. If a new service needs to be introduced, or an existing service needs to be modified, all nodes providing call control functions have to be upgraded. In a large network, such upgrades can take many months to complete. As software is being modified in a large number of nodes, the risk of major faults occurring rises significantly.

The distributed call control architecture also means that individual nodes have limited information on how to deal with calls. They are therefore restricted in their ability to provide advanced supplementary services. To illustrate this point, consider a customer who owns two sites in different geographic locations. Each site has its own telephone number and calls are routed accordingly. However, if one site has reached its threshold for receiving calls, the customer would like the overflow calls to be routed to the second site. This is difficult to achieve cost effectively in a distributed call control network because the nodes in the network are unaware of the status of the sites. Similarly, if the customer wants all calls to be routed to one site outside working hours, the changes in routing data would be...