Context, Policy and Network Management
The vision for Ambient Networks accounts for heterogeneous networks composing and cooperating, on demand and transparently, without the need for manual (pre or re)-configuration or offline negotiations between network operators. To achieve these goals, ambient networks management must become dynamic, distributed, self-managing, self-policing and autonomic responsive to the network and its ambience.
The aim of the Context, Policy and Network Management is therefore to propose integrated management solutions that support cooperating networks in highly dynamic situations where the need for autonomic behaviour is paramount to reduce complexity and operating costs. These solutions are to allow some level of network functions self-organisation and reconfiguration, namely based on tight interaction between a network-context management system, a scalable policy framework and an underlying pattern-based network monitoring system.
Because of the ability of an Ambient Network to adapt to evolving situations and conditions, new network resources can become active, policies can change and the business needs and models can vary accordingly. These changes are actuated through high-level policies that need to get appropriately translated into low-level localised actions allowing achievement of overall system goals. The Ambient Control Space is supported in such scenarios by providing on the one side constant awareness of underlying network features, resources, faults and performance and on the other the tools that allow an Ambient Network to manage and reconfigure itself.
Figure 1: Context, Policy and AN management
The innovation of this area evolves by leveraging the synergy and advances of three different research fields (network context management, network management and policy based systems) to provide a solution particularly suited for the management requirements of Ambient Networks and their Ambient Control Spaces.
AN Network management, will use pattern based monitoring to gather information (current configurations, loads, service usage, etc.) about the network status. The monitoring scales and handles network dynamicity by distributing requests in the network using self-organizing distribution patterns. One use of the gathered information will be to feed into the context management system, to enable it to provide the context information requested by other parts of the ACS as well as by users/applications/services via the ASI to optimise the mapping between service requirements, transport network resources and user preferences.
Results from the monitoring system will be evaluated against the policies that are currently active, which in part depends on context information. When the policy evaluation results in actions that require (re-)configuration of the network these will be effectuated by the AN management system. Reconfiguration possibility depends on the context information and policies.
The policy framework is meant to provide the rest of the AN project a blueprint enabling storage of as well as the dynamic change of rules defining the behaviour of an Ambient Network. The policy management task will also provide guidance to ensure policies are “well written” (i.e. reduce the likelihood of having conflicting policies within an Ambient Network). Through collaboration with the other project activities this work will also analyse how to write technology independent policies describing high-level system goals and how to translate those into low-level actions that contribute individually to the same overall stated “system goals”
Further Readings: "ContextWare Support for Network and Service Composition and Self-Adaptation"
|
