Within the next few years we will all be connected to networks for almost every purpose in our lives. Our TVs, our cars, our refrigerators, and maybe even our pets, are all going to be plugged into some network that will make our lives better. Better? Is that simpler or more complex? Either way it won't matter, it's going to be more interesting. Let's accept for a moment that soon everything will be plugged into the great network and we will all achieve digital Nirvana - who can say what the great network is? Is it the Internet? Is it some global, self-personalizing extranet? Is it some wireless broadband technology? The fact is that it is all of them and then some.

The IP paradigm has enabled datastores to be opened up to the masses and the data to be used in many new and creative ways. The Internet has made many of the recent technology strides possible because of its broad audience, and along with it, the scale to realize profits from creating things for the Internet environment. This newfound method of expression helped to finance the accelerated build-out of big bandwidth Internet backbone capacity. However, big hulking frame relays, ATM, private IP, and wireless networks have leveraged the same technology strides to exploit business applications that have increased productivity manyfold. As network providers continue to build out bandwidth of every variety at an ever-quickening pace, the large-scale network technology decisions take on a greater importance.

With the cost of building out networks dropping (I'm waiting to see dark fiber and optical switches at the Price Club), carriers and Internet companies are carefully pondering which technology to embrace. By no means is building out a business-class network a small-change affair. Price wars have driven margins to next to nothing on bandwidth so even with a reduced cost of entry there is high risk in a network build-out and the core technology decision can only heighten the risk. What if your target market doesn't embrace the technology? What if the transmission gear becomes too difficult or too costly to acquire? Each core network technology has inherent risks and rewards: ATM is great for video and quality of service, but can be pricey; Frame Relay is inexpensive and easy to use but not as efficient; IP is the common vernacular, and wireless will unleash the next generation of computing apps. It's going to be a tough call even for the smart guys. Today, on their own, none of these technologies are a cost-effective or reliable means of delivering any type of data between users. Major network providers are still investing in all of these technologies and their availability keeps the enterprise computing along certain ways of thinking.

Businesses will continue to outsource applications, and the marketplace for outsourcing these services will continue to grow at a brisk pace; however, the landscape of service providers is changing just as quickly. Network providers have made substantial investments in hosting centers, and hosting centers have made fairly substantial investments in networks. Neither has realized tremendous returns because neither is in control of the piece that lies between their disciplines, the application environment, and the ability to manage that environment as a straight-through process. These providers will need to evolve from being just facilities or network providers to being infrastructure providers delivering managed services at the network and application level backed by mission-critical SLAs.

As the enterprise opens further, inviting in partners, customers, and suppliers, networks and applications will need to work together seamlessly to deliver the same enterprise experience to all users. Each new participant will already have a network architecture that is the underpinning of its business. The success of an open enterprise will ride on new participants' experiences after becoming part of the enterprise environment. Next generation applications will be engineered with the understanding that application performance is dependent upon network performance. In order to properly manage one you must be able to manage the other. Managed network services will enable the bottlenecks to be reduced throughout the straight-through process and applications will be able to perform well over any network architecture. Real-world decisions are made on a cost/risk basis. Can you deliver an institutional financial trading application over the Internet? Sure, but you will have concerns such as security, performance, and scalability. Maybe an extranet running over ATM with a high level of guaranteed quality of service is a better answer, but it could cost more. Will this application be able to go wireless? The right answer really depends on the tools that will be used to build and deliver the application and what the user population's existing computing environment is.

All of the common communication technologies will have their place in the new world order of ubiquitous wired devices. What will make each technology successful will be its reliability and ability to interface to the other technologies in use today with a minimal amount of cost and re-engineering of the current business enterprise. The next generations of messaging and network services will make it possible for all of the popular network technologies to survive and prosper. Digital Darwinism will decide what each technology will be used for and at what cost. It will be the confluence of the technologies that will be the challenge. This is where scalable application platforms and customer-centric network services will rule the day. This will be a must for network and hosting service providers; they will need to have a scalable and flexible application platform at the core of the service model.

As we approach yet another crossroads of technology we will all do well to take a well-timed moment to decide how best our data should travel and where it will have to go to become information.