Hindawi Publishing Corporation International Journal of Reconigurable Computing Volume 2013, Article ID 141562, 17 pages http://dx.doi.org/10.1155/2013/141562 Research Article Self-Adaptive On-Chip System Based on Cross-Layer Adaptation Approach Kais Loukil, 1 Nader Ben Amor, 1 Mohamed Abid, 1 and Jean Philippe Diguet 2 1 CES Laboratory ENIS National Engineering School University of Sfax, B.P. 3038 Sfax, Tunisia 2 Lab-Sticc, CNRS/UBS University, CS 83818-29238 Brest Cedex 3, Lorient, France Correspondence should be addressed to Kais Loukil; kais loukil@yahoo.fr Received 5 April 2013; Revised 19 August 2013; Accepted 11 October 2013 Academic Editor: Jo˜ ao Cardoso Copyright © 2013 Kais Loukil et al. his is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. he emergence of mobile and battery operated multimedia systems and the diversity of supported applications mount new challenges in terms of design eiciency of these systems which must provide a maximum application quality of service (QoS) in the presence of a dynamically varying environment. hese optimization problems cannot be entirely solved at design time and some eiciency gains can be obtained at run-time by means of self-adaptivity. In this paper, we propose a new cross-layer hardware (HW)/sotware (SW) adaptation solution for embedded mobile systems. It supports application QoS under real-time and lifetime constraints via coordinated adaptation in the hardware, operating system (OS), and application layers. Our method relies on an original middleware solution used on both global and local managers. he global manager (GM) handles large, long-term variations whereas the local manager (LM) is used to guarantee real-time constraints. he GM acts in three layers whereas the LM acts in application and OS layers only. he main role of GM is to select the best coniguration for each application to meet the constraints of the system and respect the preferences of the user. he proposed approach has been applied to a 3D graphics application and successfully implemented on an Altera FPGA. 1. Introduction he popularity of the embedded systems has been increasing over the last years and new products are rapidly and con- tinually emerging [1]. hese various systems share common characteristics as multimedia applications and internet-based services. he irst main characteristic of such systems is a need for high performance capabilities for complex appli- cations. he second characteristic is mobility. hese systems must operate in varying conditions that depend on network bandwidth availability, on limited energy resources, and so forth. As such factors are known at run-time, these systems can beneit from self-adaptivity by using programmable or reconigurable architectures [2]. Self-adaptivity means that the system can modify dynamically the way services are provided according to imposed constraints using automata theory, for instance. hus, the system has several states and each of them corresponds to a set of working parameters: the number of executed applications, the HW resources used, and the allocated real-time operation system (RTOS) services [3]. Various adaptation techniques have been proposed in respect of the constraints of the system while giving a better quality of service. hese techniques can act at any of the three diferent levels: application, operating system, or hardware levels. Considering the growing complexity of the embedded systems, the support of new services, the limited energy, and the need for mobility, a new type of adaptation techniques is required. It is necessary to adopt a more global adaptation strategy which combines the previously described adaptation methods [1]. In this context, our work consists in the addition of a middleware layer, which allows a dynamic cross-layer adaptation of the system. A new approach of adaptation is proposed which is dedicated to embedded systems using two managers (global manager: GM, local manager: LM). he GM acts on the three layers in order to answer the great variations of the system constraints. It uses an oline precharacterized coniguration base which contains several HW/SW conigurations for each application. he GM selects from this base one coniguration for each application to