USALSIM: Learning and Professional Practicing in a 3D Virtual World - 2nd International Workshop on Evidence-Based Technology Enhanced Learning - page 75

Distributed training model: According to [7] “a distributed system is a collec-
tion of independent computers that give the user the impression of constituting
a coherent system”. In this model, the “Distributed System” is not a collection
of computers, but of three-dimensional objects (laboratory equipment, etc.) that
execute autonomously a series of tasks and interact among themselves in a dis-
tributed way to develop the training performed by the user. Each 3D object
(peer) communicates to the others the result of the internal operation produced
by the user's action, or by the action of another object, such that each object is
consequently able to react when the user interacts with it. In this horizontal
type of communication, all of the peers are alert to any communication from
the others. In this case, no single object coordinates or becomes the center of
communication. This allows all objects in a defined area to be continuously up-
dated so that a high volume of interaction does not produce a bottleneck or in-
activity due to the lack of information in any of the objects of the defined area.
This distributed architecture of 3D objects is based on a graph structure, that
is, on a set of objects (vertexes) joined by links (artists) representing binary
relationships among the set of elements.

•

Fig. 1 Visualization of the object relationships in a training set

This training model presents a potential problem: all objects contain most or
all of the logic of the training, meaning that an object with a high level of inte-
raction can function at a slower processing speed. On the other hand this type
of architecture allows the training to continue operating even if a node fails, as
long as it is not critical (is part of a required step).