Today, computer systems are increasingly pervasive, they are built over heterogeneous components and they offer features with complex interactions (Derdour, Dalmau, Roose & Ghoualmi Zine, 2010). MultiMedia Documents (MMDs) are specified in two different ways: future professionals have to be educated to perform the complex task of multimedia object creation, and they are also successfully used in various phases of the educational process (Cvetana, & Aleksandra, 2014). The implementation of multimedia applications is frequently made on the basis of components and services of MMDs to be adapted to the changing conditions (users, equipment, and so on.). Unfortunately, most of these adaptations are essentially based on the hardware and the user context. But, they do not consider the user's preferences. Currently, MMDs must be run on many platforms (mobile phones, PDAs, etc.). The data variety has been one of the most critical features for multimedia big data. Therefore, the way to manage and retrieve MMDs reflecting the users’ intentions in heterogeneous big data environments has become an important issue (Kehua, Wei, Mingming et al., 2015). The mix of uses and media requires the adaptation of documents to their execution context, unpredictable at the time of design document (Hai, Laborie & Roose, 2012; Laborie, Jérôme & Layaïda, 2011). In addition, the users are not all interested in the same information and they do not have the same expectations, knowledge, skills or interests. They are able to understand or to accept the services and documents in organization, content, modes of interaction and presentation, tailored to their needs and profiles.
The content used in an adaptive system can be very heterogeneous. Content servers can transmit a variety of formats and feature-rich documents. Complex content can be requested by a device which has limitations of processing, display, and so on. To ensure a good Quality of Service (QoS), the system should be able to transmit and translate this complex content to be compatible with the abilities and preferences of the end user. Two types of adaptation are generally considered: individual adaptation of multimedia objects and adaptation of the document structure or composition (Laborie et al., 2011).
The design of an architecture that adapts the multimedia content to any user in an heterogeneous web environment is a challenge. The currently proposed solutions do not address the problem of adaptation with complete architectures. They try to provide means to very specific needs such as adapting images for mobile, video transcoding, and so on. These solutions are often based on how pure programming, following closed and little flexible approaches. Therefore, it seems important to study how to make an adjustment more open and flexible and thus to cover several types within a single architecture (Lemlouma, 2004; Dietmar, & Klaus,2007).
These solutions are specific and are limited to some aspects of the environment (the size of the screen for resizing images, network throughput for transcoding video, and so on). It is due to the fact that there is no model that takes into account all the aspects of the context, nor a model for the description of terminal characteristics and of user preferences to guide a content adaptation process (Lemlouma, 2004). The consideration of the context is incomplete and does not consider all the entities that can control the transmission process of relevant contents. Then, we define an approach to identify all the adaptation elements and their characteristics, starting with the user him-self and the content, which may be requested.
If a model describing the characteristics of the environment is necessary, but it is not sufficient to answer to the requests coming from client applications. Indeed, this description must be accompanied with a technique establishing a correspondence between the different dimensions of context. This technique should be as complete as possible, considering all the possible adaptations. The difficulty is that there are no single source constraints. Indeed, each component of the network (client, network, server, and so on) can set its own constraints and the manager of QoS to find a compromise between these constraints.