Towards a Model Driven Autonomic Management System (original) (raw)
Related papers
A viable architecture for autonomic management of distributed software components
2010
Autonomic Computing is a brand of system design approaches which enable IT systems with self-management capabilities such as self-configuration, self-healing, self-protection and self-optimization. Although the field of distributed system management has achieved considerable advances, building autonomic management solutions for heterogeneous component-based systems presents five major challenges. First, component deployment and its management gets difficult with the growth of the system, because of the variety of component models with their own specifics. Second, each component framework provides its own way and interface for management creating redundancy and variety of management routines. Third, software components evolve separately which introduces problems with compatibility upon system upgrade. Forth, there are remote dependencies which are difficult to tack and this may cause unpredicted inconsistency of the system after component update. Finally, the integration of a managem...
2006
Autonomic Computing (AC), self-management based on high level guidance from humans, is increasingly gaining momentum as the way forward in designing reliable systems that hide complexity and conquer IT management costs. Effectively, AC may be viewed as Policy-Based Self-Management. The Model Driven Architecture (MDA) approach focuses on building models that can be transformed into code in an automatic manner. In this paper, we look at ways to implement Policy-Based Self-Management by means of models that can be converted to code using transformations that follow the MDA philosophy. We propose a set of UML-based models to specify autonomic and autonomous features along with the necessary procedures, based on modification and composition of models, to deploy a policy as an executing system.
An Autonomic Architecture for Legacy Systems
Third IEEE International Workshop on Engineering of Autonomic & Autonomous Systems (EASE'06), 2006
autonomic functionality into applications without user involvement is useful not only for application programmers who are building distributed systems from scratch, but also for users of legacy systems. This paper proposes a technique of injecting user code with autonomic primitives by statically analyzing the legacy code and partitioning it to manageable autonomic components. To harvest the benefit of this approach, a transparent underlying autonomic architecture is proposed which requires minimal user interaction to operate. Our goal is to provide users and programmers of such systems with an easy to use environment to work within, fulfilling the vision of autonomic computing.
Towards a model-driven architecture for autonomic systems
Proceedings. 11th IEEE International Conference and Workshop on the Engineering of Computer-Based Systems, 2004.
Agent based systems and architectures provide a firm foundation for design and development of an autonomic system. The key challenge is the selection and efficient use of effective agent architecture. A model-driven approach accommodates the underlying architecture to automate, as much as possible, the development process. The Cognitive Agent Architecture (COUGAAR) is a distributed agent architecture that provides the primary components and an implementation platform for this research. COUGAAR has been developed primarily for very large-scale, distributed applications that are characterized by hierarchical task decompositions and as such is well suited for autonomic systems. We propose a framework for the agent-based, modeldriven architecture for autonomic applications development. The framework consists of two main parts, General COUGAAR Application Model (GCAM) and General Domain Application Model (GDAM). Some COUGAAR related performance issues are also discussed.
For making software systems autonomic, it is im-portant to understand and model software-management tasks. Each such task contains typically many interactions between the administrator and the managed software system. We propose to model software-management interactions and tasks in the form of discourses between the administrator and the software system. Such discourse models are based on insights from theories of human communication. This should make them "natural" for humans to define and understand. While it may be obvious that such discourse models cover software-management interactions, we found that they may also represent major parts of the related tasks. So, these well-defined models of interactions and tasks as well as their operationalization allow their execution and automation. Based on this modeling approach, we propose a specific architecture for autonomic systems. This architecture facilitates gradual transition from human-managed towards au-tonomic systems.
Autonomic Software Systems: Developing for Self-Managing Legacy Systems
2014
Modern software systems have grown in complexity and expense, even while the cost for supporting hardware has decreased over time. Humans have a lot to do with why software is expensive, and they contribute to its cost in at least three significant areas: the maintenance and evolution of existing software, the run-time monitoring and configuration of executing software, and errors made during data entry and system configuration tasks. Software engineers seek to mitigate these costs by minimizing or removing expensive human participation in these areas where possible by adopting software and hardware approaches aimed at doing so. In this paper, we describe a commercial software engineering project where code reuse, service-oriented architecture, and self-* autonomic approaches were employed to extend the legacy enterprise system of a multi-channel vendor of musical equipment. In adopting these approaches, the developers were able to produce a highlyautomated extension to an existing system that increased the number of orders places by customers, extending the business value of that system.
Toward a new landscape of systems management in an autonomic computing environment
2003
Abstract In this paper we present IBM Tivoli Monitoring, a systems management application that displays autonomic behavior at run time, and we focus on extending it in order to encompass the design and the deployment phases of the product life cycle. We review the resource model concept, illustrate it with an example, and discuss its role throughout the product life cycle.
High-Level Modeling of Software-Management Interactions and Tasks for Autonomic Computing
For making software systems autonomic, it is important to understand,and model software-management tasks. Each such task contains typically many,interactions between,the administrator,and the software system. We propose,to model,software-management interactions and tasks in the form of a discourse between,the adminis- trator and the software system. Such discourse models,are based on insights from theories of human,communication. This should make,them “natural” for humans,to define and understand.,While it may,be obvious,that such discourse models cover software-management interactions, we found that they may,also represent major parts of the related tasks. Our well-defined models,of interactions and,tasks as well as their operationalization should facilitate their execution and automation.