Motivations for Mobile Agents
Motivation : Network Management systems (NMS) have in large part followed a centralized approach, based on the manager-agent model of standards like SNMP and CMIP. Data is collected from agents located on the devices and analyzed centrally on the manager. The centralized approach has failed to meet the challenges of today’s networks. It suffers from bottlenecks at the manager, large processing requirements for the management platform, and excessive network traffic between the manager and the numerous agents because all data is brought to the manager, involving many unnecessary transmissions. Also, it lacks the flexibility required by a heterogeneous environment and by the growth of services brought on by ATM and Customer Network Management. Operators are overwhelmed by the amount of data, by the number of alarms and events requiring attention, and by the complications of managing devices from a number of vendors.
The Promise of Mobile Intelligent Agents : The use of mobile agents affords new opportunities for the distribution of processing and control in network management. Mobile agents are migratory programs that move from one network component to another. Rather than transporting the data to a central location, mobile agents operate in the same locale as the data and return only relevant data or compiled data, thereby reducing the management traffic load on the network. Mobile agents are capable of acting autonomously to perform menial tasks or to provide intelligent support for high level tasks, thus placing the operator into a supervisory role with mobile agents as the operatives. Mobile agents are cooperative, such that agents can be assigned small low-level tasks and yet interact to achieve a higher level goal. Finally, mobility suggests that we can transport device-specific code, leading to opportunities for software version control and service customization.