Computing systems have steadily evolved into more complex, interconnected, heterogeneous entities. Ad-hoc techniques are most often used in designing them. Furthermore, researchers and designers from both academia and industry have focused on vertical approaches to emphasizing the advantages of one specific feature such as fault tolerance, security or performance. Such approaches led to very specialized computing systems and applications. Autonomic systems, as an alternative approach, can control and manage themselves automatically with minimal intervention by users or system administrators. This paper presents an autonomic framework in developing and implementing autonomic computing services and applications. Firstly, it shows how to apply this framework to autonomi-cally manage the security of networks. Then an approach is presented to develop autonomic components from existing legacy components such as software modules/applications or hardware resources (router, processor, server, etc.). Experimental evaluation of the prototype shows that the system can be programmed dynamically to enable the components to operate autonomously.