Mobile agents and framework built on mobile agents have been the key research area for the past few years. The major impedances like latency factor, abrupt disconnection in service, and minimal processing power, were solved in the mobile agent paradigm. Also, with the advent of intelligent framework of mobile agents, mobile agents were empowered with decision making powers and were able to roam the network in search of the best service provider. This further increased the efficiency of the system and reduced the system outage time. Although the system projected itself as the ideal solution to the real-world problems, it could not be implemented in commercial applications. This is attributed to the lack of sessions in the mobile agent’s environment. Predominantly in the mobile agents paradigm architecture was still the client-server architecture. In this chapter the framework has been extended to incorporate transaction capabilities to the mobile agents. This would enable them to perform a full transaction and complete a workflow. We present the scenario of Customer Relationship Management (CRM) where the framework could be put to use.
There has been a tremendous growth in the usage of mobile technologies for the past few years. Previously the mobile technologies were seen as an extended form of Internet and Intranet applications, but with the number of mobile handsets crossing the billionth mark, mobile environment have become an independent area of work. They offer several advantages like dynamic connectivity, smaller gadgets and information processing power irrespective of place and time Gray (1996). Although they seem to be a promising technology, their certain inherent qualities like low memory and latency factor restricts their usage to many real world applications (Jipping, 2002). Remote Procedure Call (RPC) provided by Java (Birrell, 1984), Network Command Language (NCL) (Meandzija, 1986), Remote Evaluation (REV) (Stamos, 1990) and SUPRA-RPC (Stoyenko, 1994) offered various techniques wherein the inherent short comings of mobile technologies could be bypassed. But all of the approaches lack a crucial feature: Coordination between various application nodes.
The novel approach of transportable programs (Gray, 1995; Cybenko, 1994) offers a promising solution for various issues raised. Transportable agents or Mobile agents, as they are called now, are autonomous programs that can migrate from one machine to another machine in the network. By migrating to the machine having the resource, the agents have the advantage of working on site where the resource is present and also use the processor’s power. This eliminates all the middleware that is required for transporting the data to the client’s site. Mobile agent’s paradigm provides an effective solution to the problem of low latency, poor interface and bad network conditions (Gray, 1996). The middleware and the communication control mechanism form the major workload in client-server architecture and by eliminating them we can build a better working environment and increase the efficiency of the system. This is so because the code as well as the state of code in execution is migrated to another machine for resuming its execution. This also eliminates the interface required for service access. The fact that there is no need for permanent connection makes it very suitable to the mobile environment. The ad hoc client-server model is overridden by the peer-peer model which matures into grid computing, where the machines can act as client or server depending on the environment. The programmer is swayed away from traditional multi-tier architecture to grid computing(Lauvset, 2001). Majority of the mobile agents architecture present in the literature lacked the feature of intelligence embedded into the mobile agents. The system designed by Cabri and Kendall (Cabri, 2002; Kendall, 1998) and Anand (Anand 2005) had agents roaming autonomously and making intelligent decisions. They could navigate and collaborate with other agents with minimal human intervention. They also learn and adapt to the environment. A detailed review of various mobile agents present is presented in the work by Anand (Anand 2005)