MCS Thesis Examination: Solmaz Jaberi

Abstract

    The growing complexity of network management has led studies to concentrate on high-level network goals to conduct networks. However, there is a wide range of context and possibilities to introduce these network goals on the one hand, and there are diverse networking technologies and devices to apply these goals on the other hand. This results in the ambiguity of network goal definitions and the complexity of applying them to the network. This thesis presents an integrated definition of intent and a study on using the Autonomic System Specification Language (ASSL) for specifying and verifying the properties of autonomic networks. This research addresses the problem of confirming the correct behavior of autonomic systems, characterized by self-management policies such as self-configuration, self-healing, and self-optimization. To tackle this concern, we propose to use the ASSL formal specification language, which provides formal expression to demonstrate desired properties of autonomic systems concisely and intuitively. The language is supported by a compiler that checks if the specified properties are consistent, and generates a Java program that simulates an autonomic system that executes according to the specifications

provided in the ASSL specification source code. We evaluate our approach on several case studies and show that ASSL can model different intent categories in a high-level abstract way. This work makes the following contributions: (i) it defines a set of intent objectives to make to reduce the vagueness of the meaning of network goals, (ii) it introduces the use of ASSL as a formal language for specifying intents in autonomic networks, (iii) it evaluates the approach on a set of intent-based case studies, and (iv) it discusses the deficiencies, implications and future work related to the use of ASSL for autonomic networks.

Examining Committee

• Dr. Sandra Cespedes (Chair) 
• Dr. Joey Paquet & William J. Atwood (Supervisor)
  
• Dr. Aiman Hanna (Examiner)
• Dr. Sandra Cespedes (Examiner)