In the context of a driven and collaborative on-line learning, multimedia resources, in particular videos, are more and more used. Due to their complex nature, there is the increasing need to manage the video contents in order to ensure a more fine-grained tracking on audio-video assets and obtain a continual feedback of student activities. In according to the ADL’s SCORM model, a video is usually assumed to be an atomic Learning Object: this assumption restricts the interaction between client and Learning Management System to a merely ON/OFF tracking process and limits the reusability of such type of contents. In our approach the video resources are considered as SCO or, more in details, as a SCO container; in this manner, each single component asset can be just a frame or a video segment in order to achieve the wanted tracking grain size. In according to the last hypothesis the main focus of the paper has been the design and implementation of an architecture for the tracking of video SCOs in web learning environment