Distributed Systems - Guide to Cloud Computing for Business and Technology Managers - page 79

3.2.1.2 Case for Distributed Objects and Components
Distributed object computing is an architecture, a paradigm, and a set of
technologies whereby objects can be distributed across a heterogeneous net-
work in such a manner that arbitrary assemblies of these objects can operate
as a single entity. Effectively, distributed objects extend the object-oriented
analysis and design methodology to the network as a whole. The concept
of an object refers to the nonincorporated software entity that is capable of
being merged with other objects to form an assembly. Once integrated into
the assembly, the object becomes a component.
The movement to distributed computing is because of five reasons:
1. Protection of expended investment
2. Gains in productivity
3. Protection of future investment
4. Enhanced utilization of assets
5. Effective interdepartmental data integration
The gains from realizing an effective means of creating, assembling, and
managing a distributed set of objects are many. Extant software (i.e., legacy
systems) can nonintrusively be incorporated into greater entities, thereby pro-
tecting existing investment. Order-of-magnitude gains in productivity can be
realized as the developer moves away from the line-at-a-time paradigm and
into the assembly, or industrial era of software development. For those objects
that were created using the techniques of the earlier era, the possibilities of
reuse are manifold, thus enabling investments future proof and optimizing
future investment. Efficiencies in deployment and execution are also realized
because the individual objects comprising an assembly (i.e., component) can
be hosted on those machines that can best be utilized by that particular com-
ponent. Finally, the various islands of technology can be bridged effectively
and selectively for optimal efficiency and effectiveness. Thus, the accounts
receivable, inventory, order, and accounts payable systems can be merged at
the data flow level without replacement or costly rework of the individual
applications.
3.2.2 Distributed Computing in the Enterprise
There are three primary approaches for realizing the distributed computing
paradigm: CORBA, COM, and RMI. Before we discuss each of them in detail,
here is an historical overview of their development.
Once networking became prevalent across academia and industry, it
became necessary to share data and resources. In the early years of dis-
tributed computing, message passing (e.g., using sockets developed in the
early 1980s) was the prevailing method of communication. This involved