Position paper (draft) for NMRG/EMANICS Meeting Mark Burgess Oslo University College What are the challenges for the next 5 years in system and network management? I suggest three things work discussing. 1) Decentralization There is still a persistent belief in the need to centralize management solutions. We have to demonstrate with a usable technology, however imperfect, that decentralized management can be made to work with no greater uncertainties or unreliabilities than centralized models. This is key to solving the challenges of both scalability and commericalization. 2) Economics and Environment (with and without money) The force driving the development of the network is commerce. Econmics is such a fundamental part of civilization that the effect of communications technology on commerce must be understood. Today we tend to think in the traditional terms of payment for services. But informal trading of services has long been practised by service providers, esp. within BGP peering relationships. Money is only one of several currencies by which parties can trade online. The economics of service provision is also a part of this picture. There is much to be gained by studying energy efficiency and heat wastage in data centres and deskstops also. 3) Convergent change management for general languages We end with a technical problem. In configuration management, there is the immunity model for reliable autonomic change management. But this model works only for primitive strings of single symbols. To make autonomic, policy based management work as a self-regulating enterprise, we need also to be able to implement the same techniques for general coded languages of the Chomsky hierarchy. Alternatively, we need custom developed operating systems with properties unlike those that we have today.