Dynamically Adapting Software Components for the Grid

International Journal of Computer Applications Technology and Research
Volume 2– Issue 4, 455 - 458, 2013
www.ijcat.com 455
Dynamically Adapting Software Components for the Grid
B. Dhanalakshmi
Aryabahata Institute Of Technology and Science
Jawaharlal Nehru Technological University
Hyderabad, India
Abstract: The surfacing of dynamic execution environments such as „grids‟ forces scientific applications to take dynamicity. Dynamic
adaptation of Grid Components in Grid Computing is a critical issue for the design of framework for dynamic adaptation towards
self-adaptable software development components for the grid. T h i s paper carries the systematic design of dynamic adaptation
framework with the effective implementation of the structure of adaptable component. i.e. incorporating the layered architecture
en viron men t with the concept of dynamicity.
Keywords: grid; adaptation; resources; entities; component specific level; component independent level.
1. INTRODUCTION
A grid [4] is a type of parallel and distributed system, which
aims at exploiting the ability to share and aggregate
distributed computational capabilities. Number of resources
may differ from processor to processor or processor to grid [4]
while performing scientific applications. Therefore, adopting
the software components for the programming model enables
security and portability on different resource infrastructures.
This adaptability framework project model can be used to
adapt the software components at run time to varying
conditions. This framework report gives the transparency of
adaptability in scientific and distributed applications by giving
the framework impact and its requirement. This gives the
ability of software to autonomously react and repair non
convenient events observed during program execution without
any intervention by the programmers.
Adaptability ensures that the application continuously
executes the best configuration depending on the actual
execution environment.
2. PROPOSED SCHEME
In this scheme, the applications involve several services like
information services, resources reservation/allocation, file
transfer and job launching and monitoring, which are
executed on different environments. Grid components may
change in processor availability, network availability,
resource sharing between applications, administration tasks,
failures etc...These environments constitute a disseminated,
heterogeneous, highly dynamic communication structure that
makes the applications as adaptive software includes different
mechanisms to modify the behavior of application or
components dynamically. This scheme suggests a layered
approach model to put together self-adaptive entities:
The central stage level defines the mandatory functionalities
for adaptive entities, while upper stage levels define the
structure to bring together primitive or composite components
adaptation [11].
With the allocated resources, we may define the best way that
is used by scientific applications modify themselves
depending on their actual execution environment. This
framework gives the ability of software to autonomously react
to and repair non convenient events observed during program
execution without any intervention by the programmers
/users. In the autonomous computing, adaptation is further
characterized as activity performed by code, acts to events,
performs suitable actions, identifies wrong behavior etc..
This paper presents a framework intended to help developers
in the task of designing dynamically adaptable components,
which puts the emphasis on an experimental evaluation of the
cost of using such a framework.
3. DYNAMIC ADAPTATION
In order to achieve an adaptation, a component needs to be
able to get information about the system (the component itself
and its environment), to make a decision according to some
optimization rules and to modify or replace some parts of its
code. Any scientific system amalgamates its modifications in
a crystal clear way for its end users. These modifications
include adapting to variable run-time conditions, masking
failures, performance measures and the evolution of scientific
application components. „Dynamic adaptation‟ [1] coats
different techniques for managing all these modifications in
the execution environment.
Dynamic adaptation [5] is classified into three dimensions
named kinds, characteristics and techniques. These
dimensions are introduced because they answer the frequent
questions of administrators and developers of the application.
This classification is the result of our investigation of existing
adaptable platforms.
For the sake of reusability, it is highly desirable to separate
the adaptation engine from the content of the component. I
capture adaptability within a framework as Dynamic
Adaptation for Components (Dynaco [1]). Associated to a
component's content, it forms a dynamically adaptable
component.
This paper presents a framework intended to help developers
in the task of designing dynamically adaptable components,

Volume 2– Issue 4, 455 - 458, 2013
www.ijcat.com 456
which puts the emphasis on an experimental evaluation of the
cost of using such a framework.
4. DYNAMIC ADAPTATION
Applications are of 2 types as, Resource-Aware: describe
resources options, select resources and then run and Dynamic
Adaptive: collect runtime information, consider/decide to
change resources, select resources and run.
A generic adaptability framework for decomposition of
adaptability in 4 steps: Observe the execution environment as
it evolves, Decide that the component should adapt, Plan how
to achieve the adaptation, Schedule and execute planned
actions.
Figure 1. Proposed Design Architecture
4.1 Architectural Design
This divides adaptability into some number of sub-
functionalities as Able to observe characteristics of the
environment in order to trigger adaptability; When a change is
detected, the framework has to decide an adaptation strategy
according to observed measures; Once a strategy has been
decided, the framework has to plan actions to implement it;
At last; planned actions have to be executed synchronously
with the execution of applicative code.
This model exhibits the functional description for the
adaptation process with the entities Decider, Planner and
Executor. The environmental changes received as events will
affect the decider and produces as a strategic plan for dynamic
adaptation.
The Planner derives list of actions from the strategy in order
to achieve the different steps of the process of adaptation. The
executor implements the different steps of adaptation to
modify the component.
Software components that are used in the adaptability
framework are separated into some number of functional
“boxes” disseminated into 3 levels as shown in figure 1. At the
functional level, the service provides an expected
implementation of the component is expected to do. If the
component was not dynamically adaptable, then it would have
the service.
The component-independent level [1] contains all mechanisms
that can be defined independently of the content of the
service functional box. The decider box is the start point of
any adaptation. It decides whether the component should be
adapted or not.
The component-specific level [1] holds the specializations of the
adaptation framework for the developers. The specified
framework consent the developer to focus the decider
for the needs of its component. It describes how decisions
can be made. The plan templates describe how the planner
can build plans depending on the requested reaction and on the
current execution environment.
4.2 Structural Design
This splits adaptability into four sub-functionalities as Able to
observe characteristics of the environment in order to trigger
adaptability; When a change is detected, the framework has to
decide an adaptation strategy [5] according to observed
measures; Once a strategy has been decided, the framework
has to plan actions to implement it;
At last; planned actions have to be executed synchronously
with the execution of applicative code.
4.3 Events Generation
The Decision making process[2] of dynamic adaptation given
by the policy procedure which will give the information about
the decisions and strategies that are used to change the
component‟s behavior depending on the execution
environment.
Monitors [13] are the entities that are used to create the events
and these events are helpful in the monitoring of components
execution. All these events are observed by decider or
monitor.
4.4 Adapting the Components
Subsequent to adaptation plan, next is the executor‟s turn,
which is regarded as a virtual machine [13] that will monitor
the control flow instructions with in the adaptation plan
depending on the execution schedule. To do so executor
depends on the adaptation points [2] this will have the
information about component states. The component states are
constrained by integrity and consistency requirements.
Observe
Decide
Plan
Schedule

Volume 2– Issue 4, 455 - 458, 2013
www.ijcat.com 457
Figure 2. Basic Design Principle
4.5 Algorithms
Following are the two algorithms that are used in this
framework.
4.5.1 Converter
The first algorithm, the converter converts a given component
to particular interface by deterministically matching the given
component/interface pair to an adapter on demand. The
algorithm will rely on the executable notations for interfaces,
adapters and components, e.g., it will have to find out at
runtime which edges a given component supports and which
borders a given adapter maps to.
4.5.2 Binder
The second algorithm, the binder combines the given
interface with an implementation instance, which is
considered as a component or a component enclosed by
adaptors.
4.6 Autonomic Implementation of the
Schema
The schema notation of this framework is as below.
do
{
decide:
what has to adapt
what has to discard
how to adapt
on trigger:
decide possible implementations for the policy
commit:
plan for predefined implementation
plan adaptive code mechanisms
schedule all the executable operations
} while(!end)
Monitors (push)
Monitors (pull)
Figure 3. Structure of an adaptable component using this
Framework [13]
4.7 Realization as a Framework Model
In the Framework model realization, the two contents are
defined as: The content description realizes the component
state functionalities; the membrane [2] is a possessor
consisting of non-functional services that manage the
component‟s behavior. In an adaptation plan, dynamic
execution environments depend on the behavior of
dynamically adaptable components. The Control Manager is
responsible for the existed components. A modification
controller (mc) realizes the adaptable component‟s actions.
The executor [13] execute actions depending on the plans
given by the planner. Server sides (push model) and client
Decid
er
Policy Guid
e
Event Strate
gy
Plan
Plann
er
Execu
tor
Observe
r
Planner Executor
Control Manager
Monitor Controller
Content Description
Server Side Client Side
mc mc mc
Membrane

Volume 2– Issue 4, 455 - 458, 2013
www.ijcat.com 458
sides (pull model) are applicable to the adaptable component
while executing them.
Modifications are possible by pushing and pulling the
adaptation methods by which the model implements the
adaptability concept on components.
5. CONCLUSION
This framework analyses how to design dynamic
adaptability support for scientific applications. It is
independent of formalisms and technologies. Also
evaluate the proposed model as well as the possibilities
to write generic adaptation policies at the collection and
application levels.
This framework for adaptability is independent of the
applications like numerical algorithms, transaction
systems etc…
Still lot of problems to be investigated / solved
(adaptation policies, performance models ...). For the
future work, the activity of generalizing the approach is
considered. i.e., generic definition of global adaptation
points should be implemented.
6. ACKNOWLEDGMENTS
My sincere thanks to the experts, who have contributed
towards development of this framework.
7. REFERENCES
[1] Je´re´my Buisson, Françoise Andre´, and Jean-Louis
Pazat. Dynamic adaptation for grid computing. In
P.M.A. Sloot, A.G. Hoekstra, T. Priol, A. Reinefeld,
and M. Bubak, editors, Advances in Grid
Computing- EGC 2005 (European Grid
Conference, Amsterdam, The Netherlands,
February 14-16, 2005, Revised Selected Papers),
volume 3470 of LNCS, pages 538–547,
Amsterdam, February 2005. Springer-Verlag.
[2] Je´re´my Buisson, Françoise Andre´, and Jean-
Louis Pazat. Performance and practicability of
dynamic adaptation of parallel computing : an
experience feedback from Dynaco. Publication
interne 2006 Projeeects Paris.
[3] Je´re´my Buisson, Françoise Andre´, and Jean-Louis
Pazat. Enforcing consistency during the adaptation
of a parallel component. In The 4th International
Symposium on Parallel and Distributed
Computing, July 2005.Eason, B. Noble, and I. N.
Sneddon, “On certain integrals of Lipschitz-Hankel
type involving products of Bessel functions,” Phil.
Trans. Roy. Soc. London, vol. A247, pp. 529–551,
April 1955.
[4] Greg Burns, Raja Daoud, and James Vaigl. LAM:
An Open Cluster Environment for MPI. In
Proceed- ings of Supercomputing Symposium,
pages 379–386, 1994.
[5] Introduction to Grid Computing – A IBM‟s red
book for details about Grid Computing is also
useful for installation of Globus Tool Kit 4.
[6] Jérémy Buisson, Françoise André and Jean-Louis
Pazat. A framework for dynamic adaptation of
parallel components. In ParCo 2005, Málaga,
Spain, 13-16 September 2005.
[7] Pierre-Charles David and Thomas Ledoux.
Towards a framework for self-adaptive
component-based applications. In DAIS’03, volume
2893 of LNCS. Springer-Verlag, November 2003.
[8] Brian Ensink and Vikram Adve. Coordinating
adaptations in distributed systems. In 24th
International Conference on Distributed
Computing Systems, pages 446–455, March 2004.
[9] Brian Ensink, Joel Stanley, and Vikram Adve.
Program control language: a programming
language for adaptive distributed applications.
Journal of Parallel and Distributed Computing,
63(11):1082–1104, November 2003.
[10] Introduction to Grid Computing – A IBM‟s red
book for details about Grid Computing is also
useful for installation of Globus Tool Kit 4.
[11] Research group on “Performance models and
adaptivity”:
http://www.di.unipi.it/~marcod/WP3homepage/RG
_adaptivity/index.html
[12] Segarra, M.T. ; Dept. of Comput. Sci.,
IT/TELECOM-Bretagne, Brest ; Andre,
F.Autonomic and Autonomous Systems, 2009.
ICAS '09. Fifth International Conference on
« Building a context-aware ambient assisted living
application using a self adaptive distributed model
[13] http://hal.archives-
ouvertes.fr/docs/00/05/76/49/PDF/PI-1782.pdf
.

Dynamically Adapting Software Components for the Grid

More Related Content

What's hot (20)

Similar to Dynamically Adapting Software Components for the Grid (20)

More from Editor IJCATR (20)

Recently uploaded (20)

Dynamically Adapting Software Components for the Grid