Real-Time Systems Laboratory Seolyoung, Jeong The CASCADAS Framework for Autonomic Communications Autonomic Communication Springer

Introduction CASCADAS Framework –ACE Component Model Semantic Self-Organization Situation-Awareness Pervasive Supervision Security and Self-Preservation Pervasive Behavioral Advertisement Scenario Conclusions Contents

CASCADAS (Component-ware for Autonomic Situation-aware Communications, And Dynamically Adaptable Services) 3 year project from 2006 – 2008 Consortium –IT (Telecom Italia) –BT (British Telecommunications), UK –12 partners from Academia and industry Goals –design, develop and validate a distributed framework (based on ACE) for composition and execution of situation-aware and dynamically adaptable communication services. –cope with dynamic and uncertain environments Introduction

CASCADAS Framework Fig. Architecture of the CASCADAS Framework considers developing and deploying applications and services in terms of ACE components or ACE aggregates. ACE-based tools to enforce specific properties concerned with actual network architectures and with physical sensing and embedded systems

ACE (Autonomic Communication Element) ACE Component Model Fig. State diagram of ACE’s lifecycle management Manager is in charge of lifecycle and internal event handling Supervision Bus and Gateway Checker Objects observing and controlling inter/intra communication events of an ACE Gateway for external communications Executor is the organ in charge of executing the plans Plan contains a set of states and transitions which have to be processed by the Executor Functionality Repository enables specific functionalities to be deployed and accessed on request Self Model is a set of predefined plans (in XML) representing the potential behaviors (e.g. in terms of invoked functionalities) of an ACE. Facilitator analyses it and uses its rules to continuously generate and/or update Plans. create/modify read execute call functionalities

Supports through algorithms for organized self-aggregation of ACEs. –clustering –differentiation –synchronization Two parallel approaches for introducing self-* algorithms: –extending the gateway –creating new self-models Semantic Self-Organization Fig. Self-organized aggregation through local rewiring with three ACEs. Fig. rewiring algorithm

Situation-Awareness Fig. Architecture of the KN KN is in charge of gathering and processing information to form a collection of Knowledge Atoms (KAs) KAs data model was created to represent any fact, in a simple and expressive way, by means of a 4-tuple of the form (Who, What, Where, When). This 4-tuple represents the basic unit of information in KAs and allows to account an entity (Who) involved in some activity (What) at a certain location (Where) at a certain moment (When). takes from context-awareness. advances with techniques to organize the amount of distributed information in proper, strongly distributed “knowledge networks” to support situated and adaptive service provisioning.

refers to the ongoing observation of ACEs and the issuing of corrective measures upon detection of hazardous situations Pervasive Supervision Fig. ACEs for pervasive supervision

A distinctive feature of ACEs framework is the absence of a centralized authority. A-priori trust relationships between ACEs belonging to different administrative domains cannot be assumed. –ACEs can show selfish, uncooperative or, in the worst-case, malicious behavior. Therefore, it becomes of paramount importance to address security issues in two distinct directions to cope with this wide range of attacks. Security and Self-Preservation

Pervasive Behavioral Advertisement Scenario Fig. The Pervasive Behavioral Advertisement scenario

ACE conceptualization and abstract model Development of pervasive supervision functionalities Evaluation and development of algorithms for automated aggregation Evaluation and development of self-preservation techniques for ACEs Identification of models for the organization, correlation, and composition of knowledge networks Quantitative evaluation of autonomic capabilities. Geographically distributed test-beds Conclusions

Thank you!