• Downsizing is enabled by a number of factors:
• Powerful cost effective PC platforms.
• Mature networking technology providing a communication infrastructure.
• Software that delivers a robust development environment for the creation of applications with mainframe-like complexity.
• Economical migration paths for moving applications to the desktop.
• The ability to integrate and synchronize front-end tools and back-end data sources.

Downsizing is also enabled by the emergence of the Cooperative Processing model. Cooperative Processing is an approach to computing in which systems are designed to function in a modular fashion. In effect, Cooperative Processing allows intelligent PC workstations to access data from throughout the enterprise; that is, on multiple database management systems residing on multiple platforms.

Downsizing has emerged as the key trend among enterprises worldwide in the 1990s. The emergence of PC-based LAN'S, the rapid application development offered by the PC, and the PC's affordability all support the argument that downsizing and the Cooperative Processing model will be the dominant approach to managing information systems throughout the 1990s .

This white paper provides Netmation's perspective on why downsizing is a sensible, affordable and adaptable approach to managing information systems into the next century. Or, in short, why the architecture that enables downsizing is truly the one to build on.

A Historical Review
To get a sense of what lies ahead, it's often helpful to take a look back. For information systems, this look back shows an evolution from mainframe-dominated systems, to PC-based approaches, and then to today: A balance between the two.

In the 1970s, the prevalent model for information management systems was a centralized hierarchical database platform. Such databases, residing on a mainframe or minicomputer, offered high performance and tight security but little if any end-user access or control. Essentially, the mainframe provided a highly sophisticated server with unintelligent clients, or "dumb" workstations. For its day, the approach was sound, giving information system managers a way to electronically store huge amounts of information that previously filled volumes of books and ledgers in company storage rooms.

But it soon became clear that accessing and manipulating these databases was difficult, as well as prohibitively expensive. The 1980s saw the introduction of PCs that were used by individuals and departments within an enterprise to provide local, easy-to-use decision-making tools. Over time these users became more sophisticated in their usage, their expectations rose and they had a desire to communicate with others in their group and to have greater access to all corporate data. The solution at the time was to link these PCs through Local Area Networks, giving the users increased information access.

This approach, which used a "file server" model in which the server acted as an unintelligent storage device, did offer peripheral sharing but, there was no intelligent data management capability, which severely limited the types and sizes of applications that could be used. The immaturity of this approach was reflected in the lack of well integrated and transparent connectivity to mainframe and minicomputer systems, where much of the corporate data resided (and still remains). In short, workstations were restricted in their ability to handle complex applications and to get at relevant data.

Information system managers were highly interested supporting their end-user's desires and in gaining improved access to all their corporate databases through the use of PC technology. A significant concern of these managers though was a guarantee that they could maintain security, integrity of data and accessibility. The existing single platform approach would no longer work, a new model of computing architecture was needed.

The Cooperative Processing Model
Since the late 1980s, one of the dominant influences on the downsizing trend is the ability to connect the capabilities of PCs linked in local or enterprise-wide networks with mainframes and minicomputers, attempting to utilize the best of both worlds. This approach, Cooperative Processing, could not be fully accomplished before, simply because the technologies that allowed PCs seamless and transparent access to heterogeneous data formats and platforms were not available. Certain things had to happen before Cooperative Processing could emerge as a viable approach for developers.

What was needed to accomplish true cooperative processing is technology giving developers the capability to transparently access databases residing on multiple platforms throughout an organization and allow the sharing of information among these locations. This ability would permit intelligent workstations to directly read and update information in any data format, from any data manager, spreadsheet, or other application, whether that information is across a LAN, on a database server, minicomputer or mainframe.

Being able to seamlessly access disparate information systems in an enterprise would allow the merging of the immense amount of information stored in databases.

In doing so, a company's mission-critical information residing in multiple databases and platforms would be brought to the desktop -that is, downsized. In downsizing, developers take advantage of the economies of the PC environment, delivering user-friendly, customized applications that can be developed more rapidly than before. The bottom line is that mission-critical applications can be brought to the desktop where they are more effective and investments in existing information systems can be saved.

This is not to say that all components of an application or information system should be brought to the desktop. Intelligent downsizing directs developers to bring that application that is best utilized on the PC to the PC; likewise, applications that work optimally at the mainframe stay at the mainframe, with access to the desktop maintained. This selective downsizing is aptly called "rightsizing", and is made possible by the modular Cooperative Processing model. In rightsizing, the mainframe and the minicomputer are not necessarily eliminated - they can be used more effectively in new roles or to run applications that do no require PC user interfaces.

Downsizing: A Critical Edge
By downsizing applications to the desktop, organizations can more easily deliver vital information to its users, providing a competitive edge. With the cooperative processing model these organizations can integrate their applications and information system resources with others throughout the enterprise. This approach, which blends the flexibility and power of the PC (linked more and more through LANS) with the benefits of the resident database management system, offers numerous advantages:

An Open Computing Environment. Despite well-intended and energetic attempts to impose a single computing standard on the industry, the truth is that corporate information systems are and will remain heterogeneous. Standards provide for a level playing field where users can invest and be assured that their investments will be protected. And, standards provide a platform for software innovation, a mainstay of the computer industry. Integrating new technologies and approaches with existing standards should be a goal of every company. The challenge in running this type of heterogeneous environment is to find the correct balance between innovation and standards. Therefore, a information system designed to work with existing processing environments, utilize industry standards and be capable of taking advantage of emerging technologies will deliver the most effective business solutions.

The desire to balance new technologies and trends with existing resources is a large part of the motivation behind downsizing. By providing a modular architecture and component pieces that easily connect, the cooperative processing model delivers an environment that enables enterprise-wide connectivity and integration of heterogeneous data formats. The alternative is to create a homogeneous environment that by its nature is limited to its native capabilities.

Rapid Application Development
Part of the attraction of downsizing is the ability to rapidly prototype, develop and implement applications. One of the principle shortcomings of a mainframe or minicomputer-based information management system is the application development time required and the resulting backlogs. Building mission critical systems at the desktop is literally months and years faster than on a mainframe or minicomputer. Having robust development tools that enable the creation and implementation of systems in weeks rather than months allows organizations to benefit more from their information and realize a significant dollar savings in the process.

Leveraging Existing Investments
Information management systems that handle mission-critical applications represent investments of millions of dollars and thousands of man-hours of work. When faced with the need to expand or upgrade these systems, scrapping the existing investment and building a new one is simply not a financial option. Nor is it necessarily a wise one. What is required is the ability to responsibility migrate these systems to the desktop by integrating the various platforms.

Development Flexibility
The purpose of any software application is to provide a service that makes accomplishing a task efficient and economical. Downsizing mainframe and minicomputer applications to the desktop offers the application developer a wide range of PC-based tools to assist them in meeting needs of specific applications. The flexible nature of this environment lets the developer work faster, maintain applications with less effort, and be more responsive to the needs of the organization.

Financial Considerations
Developing applications on the desktop platform is a much more cost effective than traditional mainframe or minicomputer approaches. PC-based development means your hardware and software investment is less, development time and its related costs are reduced and with the ability to leverage existing information system investments, and you do not incur the large expense of starting from scratch.

Other Advantages
Downsizing also offers smoother application maintenance; freedom from being locked into proprietary systems that may become outdated; and the ability to scale applications to match the changing needs of the organization. Ultimately, downsizing gives users improved tools and critically needed access to data, which in turn increases productivity and the effectiveness of an organization.