An agent running on the system is capable of local data collection, correlation and processing. Taking PC power management as an example, the agent can make better decisions based on activity that happens locally, for example whether the user is active before prompting to power off the system.
An agent running on the machine can query the operating system to check when the user last used the machine and whether he is logged on locally or remotely in order to defer or force the low power state. With multiple users logged on, each user’s documents can be saved before logging off. In summary, user productivity is not disrupted.
With agentless technology, there is reliance on remote methods to find interactive user sessions which have a dependency on specific remote accessible APIs that cannot return whether the sessions are really active i.e. user logged on and working.
Missed a post in this series? Here’s links to some key factors that may make you think twice before listening to an “agentless” vendor.