Testing mobile agent platforms over the air

Mobile agents are considered a suitable technology to develop applications for wireless environments with limited communication capabilities. Thus, they offer interesting advantages compared with a traditional client/server approach, derived from their autonomy and capability to move to remote computers. An important benefit is that mobile agents can carry the computation wherever it is needed at that moment. For example, instead of communicating a large amount of data from a computer to a mobile device, a mobile agent can move to that computer to process the data there, saving wireless communications. Similarly, a mobile device can assign a task to a mobile agent which travels to a fixed computer with the appropriate resources, relieving the overload of the mobile device. However, most experimental evaluations of the mobile agent technology focus on distributed environments with fixed computers connected through a wired network. Therefore, it is not clear how well prepared the current implementations of mobile agent platforms are to deal with unreliable, wireless communications, and with mobile devices with limited capabilities. In this paper, we evaluate experimentally the behavior of some platforms in such contexts, with the goal of motivating further research.