In the first 35 years IBM dominated information management. When IBM mainframes became connected with terminals all logic and communication control remained under the control of IBM. All that the CIOs needed was a reliable dial tone and inexpensive bandwidth. As result, IBM earned over 50% of total industry profits.

The mainframe-centered communication model became obsolete when most of the control over transactions shifted to the Microsoft desktop. Microsoft now earns over 50% of the total profits in the industry.

What do the IBM and Microsoft models have in common? What insights can we gain to prepare us for leadership in next 20 years?

It is a fact that profits will always migrate to what I call the "universal system integrator". I define the "universal system integrator" as an organization that materially reduces the costs of using information technologies.

IBM managed to be the "universal system integrator" for 35 years. Their approach was: "do it my way". Software, peripherals and consultants had to adapt to that. It is what I label as "one-to-one" integration. The telecomm carriers were squeezed out of integration and started sliding to their current status as commodity suppliers.

Microsoft has been a "universal system integrator" for 15 years. They earned this label when the variety of software grew beyond the capacity of the IBM model to handle it. The Microsoft solution was to perfect the "many-to-one" approach to integration. The Microsoft approach - "do it my way" - is based on its hold on the desktop operating system. If you wished to offer a service or an application, you better made sure that it fit Windows rules. The carriers as well as the mainframe makers were now squeezed out of integration value-added profits and could charge only commodity prices.

We are now arriving in an era that I characterize as the "many-to-many" interaction environment. There are millions of possible sources of information. There are hundreds of millions of possible destinations (when you include mobile computing and embedded devices). There is no way how any one vendor, any one operating system or any single solution can dominate value-added integration profits from such transactions.

If you accept my view that the greatest value from information technologies is extracted through integration, then you will conclude that the most economic solution for managing billions of interconnections per hour can be only satisfied by networks that also function as transaction integrators and not just as transmission pipes.

The Federal government should now proceed with the upgrading its networks to include integration services. The current network offerings (inexpensive bandwidth plus reliable transmissions) should be augmented by high-value added integration services. Here are some of the immediate priority tasks for the new Federal Intelligent Telecommunication Services (ITS):

1. Authentication: The secure, reliable and rapid authentication of the sources and destinations of transactions has enormous value. There are costly and perhaps even technically feasible ways of delivering assured cyber-identity. After examining the economics of what needs to be done I am convinced that a centrally managed ITS cyber-authentication service is a necessity for the Federal government.

2. Message Registry: Assuring smooth interoperability of applications within a rapidly changing diversity of equipment has enormous value. The current approach for dealing with this issue is to employ consultants to apply the integration glue as patches, middleware or conversion processes. There are insufficient billable hours available to get government-wide integration accomplished by labor alone. A centrally managed ITS messaging registry is essential for interoperability assurance. It should start with a government-wide XML registry.

3. Software Configuration: Each of the thousands of government units that cultivate their own islands of automation now update software releases, apply bug fixes and invent their own unique incompatibilities with everyone else. That is expensive, unsafe and unsatisfactory. Central management of security and reliability software assurance services is a high value opportunity. ITS should offer that.

4. Web Services: There are thousands of amateur-conceived and poorly administered web pages in the Federal government. A web administration service will offer great savings, security assurance and improved E-Gov capabilities.

5. Billing Aggregation: The current practice of distributing and then reconciling detailed (consumer-like) invoices from suppliers to users is enormously wasteful and does not add value. In a number of cases the administrative costs of billing and auditing exceed the costs of the service that has been provided! Simplified subscription billing that aggregate the charges from many vendors into a "per presence" flat monthly rate will increase the utility of the ITS.

To sum up:

• The era of the mainframe-centric architecture (the IBM one-to-one solution) is gone.

• The era of the desktop-centric architecture (the Microsoft many-to-one solution) is on its way out.

• The era of the network-centric architecture (the many-to-many solution) is on its way in. The future Network Services organization should enable the Federal Government to become a leader in how to leverage information technologies for productivity and effectiveness by adopting a more expansive view of services it will deliver.