| Today, telecommunication networks are expected to provide robust broadband transport for a wide variety of distributed multimedia applications. Regrettably, this comes at the price of, among other things, an increased complexity in call-setup network signaling. The increase in complexity is the most significant obstacle in the evolution of networking, giving rise to long and involving signaling operations.; I submit that the signaling complexity involved in call-setup can be significantly reduced by augmenting existing signaling architectures with mobile agent facilities. The use of mobile agents provides two vital advantages toward the control of complexity in a signaling system. First, agent-augmented signaling can be used to implement centralized call control across multiple service providers. Using agents, both users' call behavior and providers' service provisioning can be completely encapsulated and controlled within the network management system. This permits user-specific service creation and deployment, also known as open signaling, maximizes network's flexibility. In this dissertation, I demonstrate the ability for agent-augmented signaling to be used to prototype and deploy new network services. Second, agent-augmented signaling permits complex signaling operations to be implemented with minimum number of transactions between signaling entities, and thereby reducing latency associated with this operation. This reduction enables nontrivial operations to be performed during the short interval allocated for performing call-setup. The agent-augmented signaling facility forms a “division-of-labor” relationship with the resident network signaling infrastructure. Complex operations that suffer lengthy signaling delays are performed with agents while all other operations are performed within the resident infrastructure. The ability to perform complex signaling operations permits quality of service (QoS) negotiation and resource allocation optimization to be performed during call-setup. In this dissertation, several negotiation policies are investigated for their suitability to attain best deals in a call-setup session. This dissertation also proposes a negotiation mechanism used to permit structured cooperation among negotiation participants, allowing compromises to be made. Finally, this dissertation proposes a QoS optimization guideline, providing a way to specify optimization algorithms that are aimed to maximize the utility derived from the service.; A proof of concept is presented in this dissertation. A simple agent-augmented signaling infrastructure is implemented and commanded to handle different switches and/or routers, as well as users with varying demands. In the simulations, I demonstrate that open signaling can be implemented using the agent-augmented signaling framework. In addition, it greatly reduces latency for sophisticated signaling operations. |
