Title :
Programming systems for autonomy
Author :
Konstantinou, Alexander V. ; Yemini, Yechiam
Author_Institution :
Dept. of Comput. Sci., Columbia Univ., New York, NY, USA
fDate :
6/25/2003 12:00:00 AM
Abstract :
This paper describes a new approach to programming autonomic systems. Autonomic functions are integrated into element objects at design time using a special language called JSpoon. JSpoon extends element classes with management attributes representing configuration, performance, status and fault information. The JSpoon compiler generates respective code and interfaces to instrument the data in a common modeler repository, provided by NESTOR (Yemini at al., 2000) JSpoon programs access and manipulate management data without distinction between "agent" and "manager" roles. JSpoon further supports integration of plug-in knowledge modules that can interpret and control element operations. These knowledge modules are used to incorporate autonomic operations with elements. This design-time approach offers several substantive advantages over current alternatives. Management is integrated with the element development life-cycle. Instrumentation is compiler-generated and may be flexibly designed by element developers, while being consolidated into a unified global management data model. Knowledge modules can be seamlessly integrated with third party elements augmenting these elements with the logic for autonomic behavior.
Keywords :
adaptive systems; configuration management; fault tolerant computing; knowledge representation; middleware; object-oriented programming; program compilers; JSpoon compiler; JSpoon language; NESTOR; autonomic behavior; autonomic function; autonomic system programming; code generation; compiler-generated instrumentation; configuration management; element class; element development life-cycle; element object; fault information; interface generation; management data model; management management; modeler repository; performance management; plug-in knowledge module; status management; Computer architecture; Computer science; Conferences; Costs; Data models; Information management; Instruments; Knowledge management; Logic; Runtime environment;
Conference_Titel :
Autonomic Computing Workshop. 2003. Proceedings of the
Print_ISBN :
0-7695-1983-0
DOI :
10.1109/ACW.2003.1210219