Autonomic Computing: A Revolutionary Paradigm
for Implementing Self-managing Systems
Pradeep Kumar Singh, Arun Sharma, Amit Kumar, Ayush Saxena
Department of Computer Science & Engineering
Krishna Institute of Engineering & Technology
Ghaziabad, India
arunsharma@ieee.org , pksingh@kiet.edu
Abstract—Recent developments in the field of Information &
Communication Technology (ICT) have led to a large scale
installation of computing devices connected to each other via
different networking technologies. But the scale, complexity,
dynamism and heterogeneity of these devices have rendered
networks, systems & applications insecure, unmanageable &
brittle. Soon the number of these devices will outnumber the
humans using them, which makes managing them a cumbersome
task. The level of expertise needed to handle these complex
systems will also raise. Hence there is a need of alternate system
and application design paradigm based on strategies used by
biological systems, especially Autonomic Nervous System (ANS).
Based on ANS, self-managing computer systems and software
tools are being realized. Various subsystems of ANS such as
immune system may also motivate growth in developing self-
protecting systems. This paper presents a study on autonomic
computing along with its architecture, metrics, design issues and
some applications.
Keywords: Autonomic Nervous System.
I. INTRODUCTION
The recent developments in computing, communication and
information infrastructure have resulted in an explosive
growth in computer systems and applications that pervade
every aspect of our life. Soon the number of these devices will
outnumber the humans using them, which makes managing
them a cumbersome task. Therefore scale, complexity,
dynamism and heterogeneity of these devices have rendered
networks, systems & applications insecure, unmanageable &
brittle. The level of expertise needed to handle these complex
systems will also raise. Even IBM cited that applications and
environments that weigh in tens millions lines of code will
require skilled IT professionals to install, configure, tune, and
maintain the system [1]. Two solutions to this are either use a
large number of intermediate managing devices or use
intelligent devices capable of making decisions and managing
themselves according to workload. Out of these the former
will attribute to increase in complexity but the latter is a fine
solution. One such way is to take inspiration from the
biological systems. A closer look at the autonomic nervous
system (ANS) in humans has revealed that it frees our
conscious brain by controlling our heart rate and body
temperature. This methodology can be incorporated in
distributed systems so that the main controllers (human) need
not to worry about the small bugs and problems. System must
be capable of sorting out as many problems as possible itself.
Thus we require a system which could ease the management
of operations autonomously. Autonomic Computing is a big
step in this direction. Like ANS, Autonomic Computing also
hides the inner complexity providing view of an autonomic,
ubiquitous computing environment that provides a level of
abstraction and an interface which satisfy user needs. The
interaction mechanism between the physical layers of
hardware is not known to the user. The system can make
decisions spontaneously on its own, using its knowledge base
and high level guidance from human user. This system will
constantly monitor the environment that affect it and adapt
accordingly so as to optimize the performance and also protect
as well as recover from problems. Current computing
paradigms are based on static requirements, behaviors,
interactions and composition. But autonomic computing
paradigm can effectively manage dynamic nature of growing
ICT infrastructure. When, operating systems came into
existence they provided a platform to run application programs
thus making user free of inner complexity of memory
management, inter-process communication, resource
allocation, file management, etc. Autonomic systems will thus
make advancement over operating systems. They will create
an extra functionality layer over network connected devices
each running their own OS. Thus as OS manages and
coordinates individual components, Autonomic systems will
manage and coordinate distributed, heterogeneous systems.
Autonomic computing requires scientific and technological
advances in many fields in order to meet its objectives. Thus,
new architectures are required to support effective integration
of constituent technologies. However journey towards fully
autonomic system is quite distant in terms of time and
technology.
II. AUTONOMIC NERVOUS SYSTEM
Autonomic Nervous System (ANS) is the body’s master
controller. ANS is subdivided into sympathetic and
parasympathetic which counteracts their effects. ANS works
without any control from human. Some functions of ANS are
regulation of blood circulation, maintenance of blood-glucose
concentration, protective reflex on heat contact, sweating to
maintain body’s temperature, hormone secretion etc. Using
Ashby’s ultra stable system, it can be seen that there may be
2011 International Conference on Recent Trends in Information Systems
7
978-1-4577-0792-6/11/$26.00 ©2011 IEEE