E-volution of Computing Architecture and Applications

Phase I: Mainframe and legacy computing

It all started in the 1960s with mainframe and legacy computing. The

mainframe model is a fully centralized architecture in which a central system

performs all of the processing and contains all of the data. Users access this

system through ‘dumb terminals’.

As legacy systems e-volved, we saw the transition into distributed

minicomputer architecture, which is a medium-scale computer that functions as a

multi-user system for up to several hundred users. The minicomputer era also saw

organizations focus on the internal functions of the enterprise.

Indian IT vendors were very hardware-centric in those days, offering limited

Unix-based coding engagements to organizations that were largely focused on

internal functions such as selling, building, or delivering the product or

service.

Companies such as Tata Consultancy Services (TCS) and HCL Infosystems were

two of the few Indian names, which focused on hardware and hardware enabling

tools during this era.

Legacy systems demand very batch-oriented processing codes. These

applications were command-specific, rigid, and tailored to the device which they

were driving. Procedure-based programming languages in vogue during this era

included UNIX, PASCAL, and COBOL.

Phase II: Client-server

The next large-scale paradigm shift gained momentum in the 1980s with the

introduction of client-server technologies where the client (PC) is the

requesting machine and the server is the supplying machine, both of which are

connected via a local area network (LAN).

IT services demand boomed for migration work into this latest era of

computing–one which still dominates the corporate enterprise today.

The global demand surge for migration work and client-server applications saw

a major boom in Indian IT service providers, and the onset of a hardware-centric

focus to one based on services, reengineering, and application development.

Companies such as Infosys Technologies, founded in 1981, were ushered in by this

era.

The migration to client-server architecture brought with it an onslaught of

creative, object-oriented programming languages. C, C++, SQL Server, and

Microsoft’s Visual Basic introduced the use of more modular and flexible

programming. While a big improvement from procedure-based command-specific code,

these applications were still largely device-specific.

Phase III: E-volution driven by the Internet

As client-server architecture evolved, servers took on more of the processing

of applications. The Internet age and thin-client architecture emerged, and we

find ourselves back in an era of powerful servers dedicated to running entire

processes.

Thin client architecture, true to its name, implies that virtually all heavy

processing will migrate from the user’s (or client’s) desktop to more

powerful Web servers–leaving the client-end thin and opening the path to

device neutrality or the use of personal digital assistants (PDA) and mobile

phones.

Internet-enabled thin-client architecture is influencing organizations to

reorganize themselves around business processes such as order fulfillment and

product development, reducing the importance of functional boundaries. Embracing

the e-volution means that businesses need to ‘Webify’ their existing

platforms and business processes to facilitate any-to-any connectivity.

Internet-based thin-client applications are open platform-based and extremely

flexible. Using the Internet as a platform means adopting Internet standards

which can be used by anyone (i.e., they are open); hence allowing developers to

break away from proprietary standards like those of Microsoft. In this wave of

computing, programmers have started writing software for browsers instead of a

rigidly defined environment such as Windows.

These new object-oriented programming languages include COBRA/DCOM, Java,

HTML, and XML. And, as stressed above, they are device neutral–they can talk

to computers, TV sets, and PDAs.

In essence, applications enabling thin clients allow complete many-to-many

interactivity and total platform ubiquity.

Internet-based applications are differentiated. Internet-based e-business

applications differ from client-server applications in that they leverage

Internet architecture to enable organizations to electronically sell, purchase,

support, and understand their customers, suppliers, employees, and partners.

With the Internet, companies will be able to organize themselves around two

critical architectures–the intranet and the extranet. The intranet utilizes

Internet technology to facilitate internal communication. The extranet utilizes

the Internet to communicate with parties outside the organization.

From an application perspective, Internet architecture has three tiers–client,

application, and data. The client tier is browser-based and is described as ‘thin’

because little or no business logic runs on the client level. It is somewhat

analogous to the dumb terminal connected to a mainframe. The thin-client tier is

important because it enables the application to run on many different, often

non-traditional, devices including PDAs, pagers, and mobile phones, in addition

to PCs.