Joseph Kroculick, Cyhthia Hood
Assurance-based design uses an assurance case to demonstrate that a system possesses a desired characteristic such as resiliency or provides a useful service. The characteristic of a network system can be a global property such as end-to-end connectivity between communications devices. We explore how assurance cases can be used to show whether the telecommunications network infrastructure is properly configured to provide a communications service to end-users. Our paper examines knowledge models which enable the assurance case to be automated and provided to a network administrator as a conflict-checking service.
The overall dependability of the telecommunications network infrastructure is impacted by the configuration of heterogeneous distributed components that communicate to achieve a resilient network . Provisioning commands set the values of protocol settings provided to an end-user in order to realize a desired network configuration. Network configurations are typically activated by running commands at a command-line interface such as Juniper’s JUNOS  and sending them to different network components and attached protocol instances. In current telecommunications management networks, local device actions are frequently made without fully understanding the impact to an end-to-end architecture and service requirements. Furthermore technology-specific and fragmented data models attached to devices need to be set to the correct values to control complex and mission critical systems characterized by dynamic connectivity requirements.
In this paper, we propose a dependability evaluation process based on assurance cases to detect provisioning errors in implementing a network infrastructure service such as network resilience. Dependability ensures that the network infrastructure will perform an operation consistent with an agreed-upon service level.
Our dependability assessment process depends on a knowledge management approach to automate the argumentation step of assurance based design. We identify considerations in applying assurance-based design to assessing network infrastructure services in the telecommunications domain including what type of information needs to be represented in the metamodel and data transformations. For example, to demonstrate error-free provisioning, a physical implementation model needs to be transformed to a logical model, which is in the reverse direction to the transformation process of model- driven architecture from logical model to physical model. We also show how assurance-cases can be applied to detect misconfigurations of recovery protocols establish new connections through signaling across distributed components or manually provisioning of each component to provide an end-to-end path through the network.
 R. de Lemos, P. A. de Castro Guerra, and cecilia Mary Fischer Rubira, “A Fault-Tolerant Architectural Approach for Dependable Systems,” IEEE Software, pp. 80 – 87, 2006.
 M. Bushong, C. Gadecki, and A. Garret, JUNOS For Dummies. Wiley Publishing, Inc, 2008.