In today's competitive business environment it is critical that organizations know how to effectively leverage their information processing capabilities. Many organizations are unable to fully exploit their own computing power. This is often due to closed application architectures and incompatibilities between heterogeneous computing environments. Inter-application sharing of computing functions and information is often difficult to achieve. Competitive advantage in the marketplace will be the reward of those organizations that are able to transcend these barriers in a cost-effective and timely manner. The recent emergence of Distributed Object Management (DOM) presents a great opportunity for organizations seeking to maximize their computing potential.
DOM is the technology that enables objects to interact across heterogeneous platforms within a distributed processing environment. DOM has the technological thrust to become a major catalyzing force that will spearhead mainstream industry-wide utilization of object oriented technologies. In order for organizations to succeed in the implementation of distributed object systems, DOM must be considered from a strategic planning perspective. This article identifies competitive advantages of this technology and proposes ways in which the organization can transform itself to best utilize DOM.
Overview of Distributed Object Management
DOM as outlined by the Object Management Group (OMG) represents an attempt to define standards for open object interfaces. The strategic ramifications of this accomplishment are highly significant for organizations desiring to build open systems architectures. The crowning achievement of the OMG has been to develop the Common Object Request Broker Architecture (CORBA) standard. The CORBA architecture defines a language neutral interface that permits a client object to request services from a server object without requiring the client object to have specific knowledge of the location or identity of the server object. The Object Request Broker (ORB) is the CORBA defined mechanism that manages the interchange of messages between objects residing across heterogeneous platforms (Figure 1).
The CORBA standard for the ORB has already been incorporated into commercial products by vendors such as IBM, DEC, HP, SunSoft, IONA and NEC. Currently, ORBs operate across combinations of DOS, Windows, UNIX, OS/2 and Macintosh. MVS is expected to join the list this year. Language support is primarily for C, with C++, Smalltalk, COBOL and REXX to follow. Many software vendors have formed alliances with ORB vendors to release products that inter-operate with particular ORBs (IBM, Apple, Novell, WordPerfect, Tivoli, Borland, Lotus, Digitalk, et. al.). The CORBA 2 standard is being defined to permit ORBs built by different vendors to inter-communicate with each other.
Next Step in Object Technology
From an evolutionary point of view, DOM is the dialectically predictable next step in object-oriented technology. DOM enables objects to realize their fullest potential by removing a series of highly limiting physical constraints that restricted the object environment to a single language and to a single computing process. By permitting objects to be language neutral, location transparent, operating system independent, and machine independent, DOM significantly expands the domain of the object environment. Objects residing on remote and disparate platforms are made transparently accessible to each other through the services of the ORB. The vision of distinct inter-communicating objects distributed over the enterprise that fulfill specialized business functions is made realizable. DOM provides a strategic pathway to interconnect decentralized business units into an enterprise-wide network of objects. This fulfills a key objective of the Information Systems department which is to exert control over the organization's information resources without reducing the business units' legitimate need for local autonomy. Strategic applications of DOM might include concentrated access to customer account information, or parts and inventory data that is dispersed across heterogeneous platforms.
Competitive Advantages
To achieve competitive advantage in the marketplace, organizations must be able to demonstrate an ability to rapidly deploy changes to business strategies, products, human resources, operational procedures and other mission critical areas. The information systems supporting the organization's business plans must be able to rapidly and creatively change with minimal system development costs and minimal impact to end users. Not only must information be readily accessible, but business functions must also be readily accessible and modifiable. Access must not be constrained by heterogeneous computing environments. Organizations must be able to maximize use of their own system resources and information assets. The mean-time-to-market for information products must be reduced.
Critical Success Factors
DOM provides a number of critical success factors that can help organizations better realize competitive advantages over organizations that have not yet embraced this technology.
Open Application Architectures
A CORBA based application will be able to access and utilize object services and components that are architecturally diverse or that may operate on different platforms. The ORB thus opens up applications for shared access within the enterprise. CORBA provides the unifying glue that links disparate applications together. Organizations no longer need to contend with 'islands of automation' but will instead profit from a community of co-operating objects distributed across the enterprise. In fact, the notion of the application itself will change. Applications will no longer be self-contained units of vertical code. Instead, applications will become collections of distributed objects organized around a common domain of business functions. For example, multiple loan processing applications might share the same workflow, decision management, credit, audit and security object services. Loan and customer objects would also be shared, resulting in maximum leverage of computing resources. (Figure 2).
Open Application Interfaces
By maintaining a language neutral interface, DOM permits the transparent connection of computing functions coded in different languages. The ORB provides a common medium through which messages from dissimilar object environments are transformed into the receiving object's native language context. The language used to construct the server object is transparent to the client object when using the CORBA standard interface definition language (IDL). CORBA proposes that an object coded in C will be able accessible to an object coded in C++ or Smalltalk. Language neutral interface capabilities will, in time, stimulate the emergence of commodity object products in the marketplace. The availability of these products will reduce the need for organizations to engage in expensive in-house development of computing functions. The boundaries between in-house applications and vendor products would be reduced by CORBA. Interfaces are simplified and made easier to code. Collections of objects can be passed within the messages that are routed between interfacing objects. This reduces the development effort that is usually required when passing strongly typed flat data structures across interfaces.
Open Access to Objects
ORBs support the transparent access to objects distributed across the enterprise's network. Objects can be relocated to different processes and platforms with less development effort and without requiring code changes to client objects. This will reduce development expense. The ORB also ensures the location transparency of objects. With ORB technology, application development can proceed without having to be concerned with the technical complexities of inter-node and inter-process communication. ORBs insulate the application developer from having to learn and code low level networking application programming interfaces. This frees up the developer to concentrate on providing solutions to business problems. Return on development investment is maximized by the ORB.
The preponderance of redundant data and functions can be greatly reduced by CORBA. Foundation objects representing common functions e.g. security, auditing. logging, error management, help, fax, printing etc. are built once and are made accessible across the distributed object environment. Substantial development cost savings can be realized by concentrating these functions into distributed objects. For example, access to credit bureaus at a financial institution can be concentrated through a single credit server object. Previously the organization may have tolerated redundant links to credit bureaus. With CORBA, a single object can be leveraged to service many applications.
Object Control and Standardization
Quality control is enhanced by allowing the reuse of veteran objects that have already been developed, tested, and debugged by others. Risks of production failures can be substantially reduced. Functional control is also enforced by DOM resulting in enterprise-wide standardization of functions. Standardization results in lower maintenance costs, reduced complexity, reduced learning curves, and greater control. Furthermore, critical business classes such as 'customer' can be standardized across the organization resulting in standardized attributes e.g. customer id and address. New opportunities for subclassing preconstructed objects will also yield development benefits and result in rapid deployment of new business functions.
Vendor Independence
By achieving an open interface to objects, DOM reduces the organization's dependence upon any particular vendor. The CORBA standard encourages vendors to construct component based products that can be plugged in and plugged out without requiring change to the client object's code. By expanding the organization's freedom to rapidly maneuver to different vendor products, the organization gains greater leverage in ensuring that it's requirements are met. This should also stimulate greater competition among vendors, resulting in enhanced product differentiation and reduced prices.
Investment Retention
By utilizing DOM an organization can retain much of it's investment in software even after languages and architectures change in the organization. By encapsulating code within CORBA-compliant object wrappers, pre-existing code can be reused and made accessible to new systems and environments. The shelf-life of software is therefore increased. Legacy systems can thus be retained and accessed in an object-oriented manner. The inevitable conversion nightmare following corporate acquisitions and mergers can be reduced by linking incompatible systems into the ORB.
Organizational Implications
Distributed Object Management technology may trigger some of the following transformations in the way information system organizations are structured and managed. (Figure 3).
Object Administration Unit
The administration of an enterprise data model should be replaced with the administration of an enterprise object model. The organization's logical view of data entities would be superseded by a view of inherited object functions and data. An Object Administration unit would administer the class hierarchy that would represent the organization's investment in objects. An enterprise class library would be managed and maintained. Object Administration would cross-pollinate ideas about distributed object re-use across organizational units. Consulting would also be provided on object modeling and design methodology. It is important that the applications developing distributed objects work conjointly with Object Administration. The object administrator will ensure that objects under development fulfill the organization's strategic plan by ensuring compliance with standards, weeding out redundancy and recommending effective re-use of objects.
Object Services Unit
The development of common object infra-structure services would be controlled by an Object Services unit. This unit would manage the purchase or development of common object services such as database access, security, printing, fax, error handling, logging, help and other generic services. They would also be responsible for managing the deployment of the ORB technologies.
Object Control and Testing Unit
Change management to the enterprise object repository would be controlled by this group. Version control, distribution, rollback and recovery of class definitions must be supported. Management of the organization's production Interface and Implementation Repositories would also be managed by this unit. Security profiles representing levels of access to objects should also be maintained. Quality control and system testing are supported within this area as well. Efficient procedures must be developed to avoid the emergence of bureaucratic obstacles to rapid deployment of change to objects. At the same time, changes to core objects that are widely re-used must be prudently managed from an impact analysis and quality control point of view.
Specialized Division of Labor
Groups of technicians with specialized technical and business skills should be tasked to develop the foundation objects. Specifications for objects would be defined by object administrators and business oriented specialists. Knowledge of the corporate object hierarchy and business functions would be synthesized into class definitions and functions. Traditional vertical application teams would downsize. Technically oriented personnel in the business units would construct applications from the foundation objects. The emergence and growth of object oriented enabling tools such as Digitalk's Parts Technology and IBM's VisualAge, which permit end users to construct applications, further lends credence to the likelihood of the future redefinition of information systems responsibilities.
Planning for Distributed Object Management
DOM requires a complete organizational commitment from senior management to business partners and systems developers. To achieve sustainable competitive advantage and economies of scale, DOM must ultimately be deployed across the entire organizational spectrum. This should happen in a planned and organized manner by defining a long-term grand strategy that lays out the blueprint for transitioning the organization to DOM. A task force of key decision makers must be formulated to engineer the strategic planning process, educate senior management about the competitive implications of DOM and define a high level view of the enterprise's distributed object environment.
Functional strategies must be defined that specify how the information system units and business units must be re-organized and trained to best exploit DOM technology. The strategic plan should delegate initial responsibility for constructing foundation objects to specialized organizational units. For example, the Information Security group would construct security objects, the Auditing group would construct audit objects. Public interfaces to these objects would be negotiated by Object Administration. New and unanticipated objects will of course emerge from the creative initiatives of application developers. Object Administration must become aware of these objects for potential enterprise-wide reuse.
Obstacles to Widespread Acceptance
The transition to a new paradigm and a new way of thinking is not going to happen over night. DOM augurs the possibility of change across the organizational spectrum. Resistance to change can be expected. The fact that the industry is still young, that commercial ORB's have only recently been announced, and that most support tools are not yet available must make organizations cautious. Multi-use class libraries containing CORBA-compliant object components have not yet become visible in the marketplace. Transaction management of distributed objects is also not yet available in the marketplace.
To achieve enterprise-wide object management, the hardware platforms as well as the network transport facilities supporting the ORB must be capable of managing high-volume message traffic. Organizations will seriously scrutinize whether their network environments can support the growth in traffic that may result from DOM on an enterprise-wide level.
Another concern is that an object management bureaucracy might emerge that is inefficient and sluggish. This is paradoxically just what DOM is supposed to eliminate in the first place. Development changes to foundation objects will always be on the critical path. Production failure of foundation objects may have serious organizational impacts. Security risks are also higher with open access to object services.
Conclusion
With the emergence of true standards for open interfaces between objects, the industry cannot avoid the opportunity to endorse the OMG model. The opportunity cost of not exploiting a chance for the possibility of true inter-operability is enormous. The fact that over 300 organizations are members of the OMG is an encouraging sign that the marketplace will ultimately embrace the OMG standard.
Despite the obstacles, the use of an object request broker offers an organization the potential to gain a competitive advantage over organizations that do not utilize similar technologies. The ability to rapidly change applications to meet constantly changing business realities in a cost effective way is a major argument in favor of DOM. The value of transparently accessing core system and business objects that are distributed across the enterprise is another compelling virtue of DOM. Organizations proactively seeking critical success factors from strategic information systems must consider the benefits of gravitating towards object oriented technology and to Distributed Object Management. Organizations that are traditionally risk averse to new technologies must ask themselves not about the risks of engaging Distributed Object Management, but rather about the risks of not engaging Distributed Object Management.
Copyright Technium, Inc. 1994
All Rights Reserved