Investigating O:MIB-Based Distributed Active Information Model (DAIM) for Autonomics Pakawat Pupatwibul, Bahram Jozi, and Robin Braun, Faculty of Engineering and IT, University of Technology Sydney, New South Wales 2007, Australia Abstract—Technological innovations in communication networking, computing applications, and information mod- eling have played a significant role in managing complex distributed electronic systems. Autonomic Computing (AC) is a concept to deal with the over growing complexity of distributed networks; this term gives systems the ability of self-management, which mean each component in AC can adapt itself to changing conditions of the dynamic environment. In this paper we investigate a new nature- inspired Distributed Active Information Model (DAIM) to allow the local decision making process, that will essentially contribute to complex distributed network environments. Details of the DAIM structure are also described, which will hopefully address the schemes of some previous network management protocols such as Simple Network Manage- ment Protocol (SNMP), Common Information Model (CIM), and mechanism like Policy-Based Network Management. Finally, we will introduce a benchmark networking system called OpenFlow for applying the DAIM model to enhance autononmic fuctions. Index Terms—Distributed Network Management, Infor- mation Model, Management Information Base, DAIM, Self- management, Artificial Intelligece. I. I NTRODUCTION C OMPLEX electronic environment refers to a group of electronic devices connected together (wire or wireless networks) to share information and resources. Any network system needs a management protocol which performs different kinds of tasks such as operation (mon- itoring the performance of the network and detect any occurred problem as soon as possible), and maintenance (fix any occurred problem or in better word, always tries to keep network perform better). Most of the proposed network management protocols are based on the Interna- tional Organization for Standardisation (ISO) definition for management model. Performance management, Con- figuration management, Accounting management, Fault management, and Security management are five charac- teristics of this definition[1]. Performance management refers to operation task by means of collecting and processing important management information. Configu- ration management refers to monitoring and controlling the effects of any device on network performance. This characteristic can be use to search any useful information when any problem occurs. Accounting management guar- antees a fair usage of network resources by any user. Fault Pakawat Pupatwibul, e-mail: pakawat.pupatwibul@uts.edu.au. Bahram Jozi, e-mail: bahram.jozi@uts.edu.au. Robin Braun, e-mail: robin.braun@uts.edu.au. management refers to maintenance task which can detect the problem, isolate and fix it in the network. Security management controls the access level of all users to the resources of the network based on specific policies[2]. In this paper, we introduce the new information model, Distributed Active Information Model (DAIM) to allow the local decision making process, that will essentially contribute to complex distributed network environments. DAIM offers adaptation algorithms embedded with in- telligent agents and information objects to be applied to such complex systems. By adopting the DAIM model and these adaptation algorithms, managing complex sys- tems in any distributed network environment can become autonomous, adaptable, and scalable. This DAIM model can enhance objects to make their own local decisions through its active performance, and thus significantly reduce the workload of centralized decision-maker. In order to achieve the system’s goal, a large amount of distributed objects in the DAIM model needs to be highly integrated. II. NEEDS FOR NEW PROTOCOL Advanced technologies have dramatically escalated over the past few decades, especially distributed networks, and play a significant role in providing management services for large and complex networks. Using human operator as manager is not economical and also error- prone. Moreover, as the complexity of distributed system grows over time, an effective computing environment is needed to ensure good quality of network services and performance. Currently, large-scale electronic sys- tems like Wireless Sensor Networks (WSNs) or Bush fire alarm system are becoming more difficult to man- age, configure, operate, maintain, and re-structure. It is important to propose a new OSS (Operations Support System) management structure to cope with such com- plex distributed networks (systems). In order to deal with this problem, IBM introduced the term Autonomic Computing in 2001, which can create its own strategies so it is able to constantly adapt itself with dynamic conditions of environment. Autonomic computing can also free networks manager from some management tasks especially low-level tasks, and in the meantime bringing better system behaviour. In this regard, each Autonomic Computing System (ACS) should have two main capa- bilities; adapting itself quickly to dynamic environment,