Advantages of Mobile Agents

The advantages of Mobile Agents are depending on their goals or on the tasks they have to achieve. The main advantage is to work locally on a system and not to be dependent on a connection to this system. This can improve the monitoring of systems or processes, which can be done locally on a machine.

One field of use can be locally monitoring the status of a computer or a segment of a network. Instead of monitoring the segment from a special computer in a centralized way, Agents can travel to a certain segment and watch locally the status of that segment. When an error occurs, another Agent with special knowledge for this error can travel to the system and try to correct it, also, the watching Agent can alert the human administrator of the error. Since the normal monitoring process requires polling the system status every certain period of time, it uses a lot of bandwidth. Sending Agents to the target system and letting them monitor the status locally saves this bandwidth and gives at the same time the possibility to let the Agent (or another, more specialized Agent) try to correct the error on the system. Another benefit of this approach is that it will work even when there are network problems, since the Agent executes locally and is not dependent on a continuous connection. Another field of use are systems which are not easy to access by humans because of their locality. This can be, e.g. the nodes of a wide area network which are normally located in the underground and far away from their supporter.

Whenever a new version of the Agent is produced, it just travels to the target system and then replaces the old Agent there. So there is no need of locally installing any piece of software once an Agent System is installed on the machine. The same technique can also be used to distribute a software update in a large network of computers. Just by sending a Mobile Agent with the knowledge how to update the systems this task will be done without human interaction. The Agent should have enough intelligence to choose the best installing method on every machine depending on the information it gets from its environment. This method is not limited to computers, it can be used for any kind of Mobile Agents enabled devices e.g. network components like routers, switches etc.

The second main advantage of Mobile Agents is the collecting and processing of information. The collecting of information is still done in a very centralized way. This means that you have a central point which collects all information and then processes it on the central machine. This is very dependent on a good connection to all of the information providers. If the connection fails, then the result is probably unusable. With the use of Mobile Agents you get a much more rough version to collect information. Since the Agent depends on a connection just to travel to and from the system which provides the information, it does not depend very much on a good connection. The search for the specific kind of information is done locally and also the first filtering of the information. The Agent will then only carry with it the really relevant information. So the Agent can process the information on the same machine where it has collected them and return back with just the results of this process. Almost the whole process is independent of the connection. The connection can be down some time or can change its protocols or ways through the network ,without having any effects on the Agent and the information itself.