DBMS, February 1996
DBMS Online: Enterprise C/S By Judith Hurwitz

A Dominant C/S Infrastructure

Winning the battle (and the war) for a standard middleware infrastructure for client/server computing.

In my column last month I discussed the battle for control over the client/server infrastructure, and I suggested and explored two possible scenarios. In one plan, Microsoft, with its OLE-oriented middleware architecture, takes center stage; in the other plan, vertical software companies with integrated middleware become the default infrastructure. This month, I'll look at two other possible scenarios: The United Front scenario, based on a common infrastructure promoted by key systems vendors; and the Internet-as-Infrastructure scenario. Again, I'd like to hear your opinions about which of these scenarios you think will be most likely to become a reality over the next five years, and why. (Email me at Jhurwitz@hurwitz.com.)

The United Front Scenario

Probability of success: 50 percent
Because of the power vacuum in the marketplace, Microsoft could potentially win the client/server infrastructure-control battle by default. This possibility is well understood by many of the key players in the industry, including systems vendors, systems integrators, and a variety of third-party software companies. In the United Front scenario, however, these vendors would select and standardize a single infrastructure provided by one vendor. Each vendor would then add a few extensions to the infrastructure to integrate it with his or her environment. This infrastructure would incorporate certain Microsoft components, but it would not place them at center stage.

What would it take for the United Front scenario to be viable? First, the selected technology (such as an implementation of The Object Management Group's Common Object Request Broker Architecture [CORBA] combined with some sort of object/relational gateway technology) would have to be endorsed by powerful players in the business. It would not be enough for one vendor who owns a certain technology to suggest that it is a standard. Rather, there would have to be a well-organized coalition of some of the most respected companies in enterprise computing. In addition, these companies would have to ensure that a significant number of development tools, packaged applications, and system management tools would be ready to satisfy the needs and infrastructure requirements of large client/server enterprises.

What will it mean if this scenario succeeds? The first implication will be that the industry has matured to the point where it is ready to establish a standardized infrastructure based on compromise. The key players would have to be able to deliver tools to help companies implement enterprise client/server programs quickly and accurately. Second, it would mean that the industry powers successfully limit Microsoft to its current role on the desktop. It would also mean that users could stop spending time developing their infrastructures and start looking for software companies that would combine this infrastructure into horizontal and vertical tools and software packages.

The Internet-as-Infrastructure Scenario

Probability of success: 70 percent
It has become apparent over the last six months that the Internet has incredible potential to form the basis or foundation of an infrastructure. It will clearly be an important part of the emerging electronic commerce market. Today, companies are establishing external home pages to communicate product and service information to potential and existing customers. In some cases, they are also getting feedback and suggestions from customers via the Internet. Companies are also investing in internal home pages that spread internal textual data throughout the organization. Companies are looking at the Internet as a means of providing online support manuals and other text-based material that would otherwise be difficult to use.

A more interesting manifestation of the Internet is as the foundation of the client/server infrastructure. For this scenario to succeed, the client/server software industry would have to become enamored of the high market valuations and projections for Internet-related technologies. Several companies that produce tools for areas such as database access, graphical development software, security management, systems management, multimedia, and so on, will begin to tie their products into the Internet. In effect, these companies will transform the Internet from an unstructured resource with significant potential to a structured, valuable resource. The Internet will provide the directory services and messaging backbone for advertising and distributing these resources. Vendors could then tie their products tightly into the Internet, creating complex and sophisticated development environments and vertical software products that leverage the underlying, standards-based components.

Another wrinkle could give this scenario an even higher probability of success. Let's suppose that the same systems vendors that would push the Unified Front scenario would also tie the Internet infrastructure in with some key middleware components such as CORBA or DCE. The combination of these components could potentially create a client/server infrastructure that would have broad, long-term implications. Ironically, this new model of computing would have more in common with the old world of host-oriented computing than with the current world of first-generation, client/server computing. The Internet-focused infrastructure will be primarily focused on the time-sharing model discarded by the industry in the late 1970s. This model will also assume a much thinner client model than the fat-client model inspired by the PC revolution. If this Internet-as-Infrastructure scenario were to succeed, the PC would become less central to the corporate infrastructure than it is today, and a shared services model would be the new standard. The other implication is that the industry's power structure would realign so that a broader set of players, some old and some new, would occupy key roles in this distributed infrastructure evolution.

Which Scenario Will It Be?

If I had to put money on one of these scenarios, I would select the Internet-as-Infrastructure scenario, combined with a healthy dose of the United Front scenario. The Internet comes closest to a standards-based infrastructure. However, in its raw state, it cannot succeed. It will succeed only through the efforts of client/server tool providers adding software development components to the enterprise. In addition, other infrastructure components (object request brokers, distributed security services, and so on) are necessary to make the Internet scenario a success. A United Front, supported by key industry players, just might be the piece necessary to enable the creation of this client/server infrastructure.

Don't get me wrong: I don't expect any of these scenarios to happen overnight. Much work remains to be done, from security and applications management to simply creating the momentum necessary to sustain such an infrastructure. It is better for corporations that are making infrastructure decisions to push vendors in the direction that best serves their businesses. Without such a proactive approach, the winning scenario will be determined by the vendor(s) most skilled at spin control.


Judith Hurwitz is president of Hurwitz Consulting Group Inc., a consulting, publishing, and research services firm specializing in client/server development tools, client/server infrastructure, and systems management. Hurwitz Consulting Group is based in Newton, Massachusetts. You can reach Judith at 617-965-6900 or via email at Jhurwitz@hurwitz.com.
Subscribe to DBMS and Internet Systems -- It's free for qualified readers in the United States
February 1996 Table of Contents | Other Contents | Article Index | Search | Site Index | Home

DBMS and Internet Systems (http://www.dbmsmag.com)
Copyright © 1996 Miller Freeman, Inc. ALL RIGHTS RESERVED
Redistribution without permission is prohibited.
Please send questions or comments to dbms@mfi.com
Updated Monday, November 11, 1996.