CHALLENGES AND PERSPECTIVES IN THE DEPLOYMENT

OF DISTRIBUTED COMPONENTS-BASED SOFTWARE

Mariam Dibo and Noureddine Belkhatir

Laboratoire d’Informatique de Grenoble, 681, Rue de la Passerelle, BP 72, 38402 St Martin d'Hères, France

Keywords: Deployment, Components based software engineering, J2EE, CCM, .NET, D&C, Generic deployment

system, MDA, Deployment process.

Abstract: Software deployment encompasses all post-development activities that make an application operational. It

covers different activities such as packaging, installation, configuration, application start and updates. These

deployment activities on large infrastructures are more and more complex leading to different works

generally developed in an ad'hoc way and consequently specific to middleware such as for instance J2EE,

.net or CCM. Every middleware designs specific deployment mechanisms and tools. The objective of this

work is to propose a generic deployment approach independently of the target environments and to propose

necessary abstractions to describe the software to be deployed, the deployment infrastructures and the

deployment process with the identification and the organization of the activities to be carried out and the

support for its execution. Our approach is model driven and our contribution is about a generic deployment

framework.

1 INTRODUCTION

An important issue of component based software

engineering is the deployment of components in

decentralised locations in an efficient, safe and

consistent manner. The deployment life cycle

encompasses all the post-development activities of

an application which makes the software useful. It is

an important step in software life cycle, which for a

long time has been reduced to only installation.

Today, the components approach and the

distribution make deployment complex. Many

deployment tools exist but they are often built in an

ad'hoc way, specific to a technology or to an

architecture and covering partially the deployment

life cycle. This paper reviews this important domain

of software life cycle, emphasizing the pros and cons

of each deployment approach.

The challenge is to develop a generic framework

encompassing specific tools and supporting the

whole deployment process.

The rest of this paper is organized as follows: Part 2

presents related works. We split this part in two

sections. In the first section, we present approaches

developed by the industrials leading to tools and

platforms developed in an ad' hoc manner.

The second section aims at specifying high level

abstraction approaches using models, meta-models

and their transformations. Examples of these

approaches are Software Dock (Hall et al., 1999),

SOFA (Bures et al., 2006), UDCM (Hnetynka,

2005) and Dance (Edwards et al., 2004). We put in

this category the works developed by the OMG

group identified as D&C (OMG, 2006b) and based

on a model-driven approach.

In the last part we present briefly our approach

highlighting the concepts and the architecture of the

generic deployment framework we propose.

2 RELATED WORKS

2.1 Industrial Approaches

In this category, most of the existing technology of

deployment is built in an ad' hoc manner. Therefore,

every system has its own tool or its own method of

deployment covering partially the deployment life

cycle. In this category, we include industrial

proposals such as CCM (OMG, 2006a), EJB (SUN,

2009], OSGI (Alliance, 2003) or .NET (Lowy,

2001). These systems integrate some preoccupations

such as the persistence or safety but supply solutions

403

Dibo M. and Belkhatir N. (2009).

CHALLENGES AND PERSPECTIVES IN THE DEPLOYMENT OF DISTRIBUTED COMPONENTS-BASED SOFTWARE.

In Proceedings of the 11th International Conference on Enterprise Information Systems - Information Systems Analysis and Speciﬁcation, pages

403-406

DOI: 10.5220/0002009904030406

 SciTePress

which are too ad-hoc. For example the composition

of units is still not taken into account, the

programming language choice is very binding, and

communication is very dependent on the context. A

general weakness is the lack of abstraction allowing

a better re-use of components. For lack of a frame of

composition in the industrial models, it is the Glue

code which assures the communication between

components if they are distributed.

2.2 Academic Approaches

Model based systems such as D&C (OMG, 2006b),

UDCM (Hnetynka, 2005) provide expressive

abstractions to control deployment. These systems

enhance a technology transition based on models

and meta-models offering more generality and

elevating the level of abstraction. The key

capabilities are:

- Identification of the common services

separation of concerns with clear identifications

of the different models (applications, locations,

deployment planner and orchestration).

- Mapping based on MDA (OMG, 2005) a

transformational approach to the underlying

middleware.

- Automatic deployment and life cycle

management.

3 OUR APPROACH OVERVIEW

We know unquestionably that the concept of

architecture with components consists in creating

reusable entities and in developing personalized

software thanks to the appropriate assembly of these

components. In such a context, the roles of

components developers and applications composers

become clearly different. So, components developers

work out on generic components whereas the

applications composers concentrate on the

application domain in assembling and in configuring

generic components available in business. Thus, a

component adheres to a component model, which

establishes the standard for the implementation and

the interoperability of the component. In such an

environment, where the development of components

tends to be more and more independent from their

re-use, it is necessary to have a deployment machine

which will allow assembly and to distribute

applications correctly with components whatever

their implementation may be. Figure 1 below

represents the deployment process of components-

based software which is constituted by several

activities in correlation. Thus, deploying a

component based software consists in distributing

components in specific locations and in managing

the constraints of placement, dependencies and

configuration. Once deployed, a software system is

available for use.

We propose a generic deployment framework.

By generic we understand, the fact of being able to

deploy any application regardless of the

implementation technology.

(1) Application modeling, concerns the

description of software architecture. It allows to

model for every application, the various components

which constitute it. For every composite or primitive

component, we can specify its assembly, its

constraints of dependences as well as its software

and material needs.

(2) Network domain modeling, concerns the

description of the locations network. It allows to

model for every domain, the various locations it

contains. For every location, we can specify the

offered material and software resources.

(3) Enterprise thesaurus modeling, allows to

model the similarity between the concepts used in a

specific enterprise or in a business domain.

(4) Strategy modeling, allows defining the

strategy used to deploy applications.

(5) In deployment processes many concepts

(application, domain, enterprise thesaurus and

strategy) are manipulated. It processes components

placements. It generates and assures the

deployment plan consistency and supplies the

specific deployment descriptor. It executes the

plan following specifics strategies to the execution

environment.

Placement is the association of an application

component and a domain node. A placement (Ci, Nj)

is valid if and only if all component Ci constraints

are satisfied by node Nj resources.

Deployment Plan, for an application A

composed of component C1 to Ci where i> 1 and for

a domain D formed from locations L1 to Lj where j>

1, the deployment plan is all the valid placements

(Ci, Lj). The placements are fulfilled by the

deployment planner which contains all data and all

necessary strategies to make the mapping viable.

The deployment plan (PIM), will be instantiated

generating in the deployment descriptor specific to

the execution platform (PSM), for instance, in

compliance with the EJB platform.

ICEIS 2009 - International Conference on Enterprise Information Systems

404

Figure 1: Generic deployment engine.

The plan execution is the orchestration of the

plan in a defined order. As for instance download,

install, activate and reinstall. Our team has

developed an environment for dynamic reinstallation

and adaptation DYVA (Ketfi et al., 2002). DYVA

covers a part of plan execution.

4 CONCLUSIONS

Deployment becomes more and more complex when

deploying large systems on large infrastructures.

On the one hand, any ad' hoc solutions for deploying

monolithic or component based systems exist.

On the other hand, there is a new approach to

deployment technology. In recent years there have

been many works in development by academics

focusing on a new generation of systems. These

approaches enhance a technology transition. They

have shown the potential of using a model-driven

approach such as MDA.

The defined models are based on expressive and

simple abstractions so the application, the location,

the deployment process and its orchestration can be

built on top of that common foundation.

We hope that the deployment framework

architecture we propose is a contribution to this new

generation of systems.

At the moment we implemented the module of

application modeling, the module of network

domain modeling and the protocol of plan

generation by taking into account the strategy by

default. In the future works we plan to add a

strategies specification language more elaborated

and benefiting from various approaches.

REFERENCES

Alliance, OSGI Service Platform, Release 3, IOS Press,

Inc., 2003.

Bures, T., Hnetynka, P., Plasil, F., SOFA 2.0: Balancing

Advanced Features in a Hierarchical Component

Model, SERA’06, Fourth International Conference on

Software Engineering Research, Management and

Applications 09-11 Aug. 2006, p. 40 – 48.

Edwards, G., Deng, G., Schmidt, D. S., Gokhale, A.,

Natarajan, B., Model-driven Configuration and

Deployment of Component Middleware Publisher

/Subscriber Services, Proceedings of the 3rd ACM

International Conference on Generative Programming

and Component Engineering, Vancouver, CA, October

2004.

Hall, R. S., Heimbigner, D., Wolf, A. L., A cooperative

approach to support software deployment using the

software dock, Proceedings of the 21st international

conference on Software engineering, p.174-183, May

16-22, 1999, Los Angeles, California, United States.

CHALLENGES AND PERSPECTIVES IN THE DEPLOYMENT OF DISTRIBUTED COMPONENTS-BASED

SOFTWARE

405

Hnetynka, P., A model-driven environment for component

deployment, Software Engineering Research,

Management and Applications, 2005. Third ACIS

International Conference on 11-13 Aug. 2005, p. 6 –

13.

Ketfi, A., Belkhatir, N., Cunin, P.Y, Adaptation

Dynamique Concepts et Expérimentations

Proceedings of the 15th International Conference on

Software & Systems Engineering and their

Applications ICSSEA’02, Paris, France, December

2002.

Lowy, J., COM and .NET component services, O’Reilly &

Associates, 2001.

OMG, CORBA Component Model Specification

http://www.omg.org/docs/formal/06-04-01.pdf,

Version 4.0, Technical Report, April 2006.

OMG, Deployment and Configuration of Component-

based Distributed Application Specification, Version

4.0, Technical Report, April 2006.

OMG, Model Driven Architecture, OMG Document

ormsc/05-04-01, 2005.

SUN, Enterprise Java Bean, Adress = http://java.sun.com,

Path = /products/ejb/.

ICEIS 2009 - International Conference on Enterprise Information Systems

406