Hi all, Searching the NS archives did not shed much light on this topic: is there a recommended procedure to model available computational resources (CPU time) within NS nodes? I'm just wondering how one could model the following scenario in NS: a node runs several distinct Agents. All Agents must provide some feedback on the consumed CPU time and a node control unit delays incoming and/or outgoing packets accordingly on congestion (meaning one queue+delay object upstream and one downstream on any link).
>From an architectural point of view it's not clear to me how Agents (the ultimate source to decide on required CPU processing time for an application-specific packet that is sent or received) can change the behavior of a classifier within the node structure or, even worse, control a queue/delay object that is located upstreams or downstreams. Any ideas/hints where to start searching? Are there any NS examples that implement such behavior? Do you know of any work in this area? thanks in advance, br Joachim
