When Edison invented electric lighting, his company was simultaneously selling generators, switches and lightbulbs. At the time he invented it, there was no preexisting infrastructure for electricity, so he had to sell his customers every component of his direct-current electricity system. Many early electricity customers were the pioneers of their day and spent countless hours building, deploying and debugging these systems...only to find out a few years later that alternating current electricity systems would become the eventual standard.

In today's world, many years later, electricity has become ubiquitous - but not without having gone through major transformations (no pun intended). Gigantic investments in infrastructure systems had to be made to bring electricity to our doorsteps. The electricity grid has become a nationwide network and is powered by enormous centralized plants making use of exotic power sources such as nuclear, hydro and coal. The electricity network has become highly decentralized and very reliable, and has many redundant components. The business model has evolved from the selling of hardware to the selling of the electricity itself. Today, when you turn on a light, a microbilling system will automatically charge you a couple of cents more on your monthly bill. Electricity has become a service; it has become mainstream.

The Internet is only a few years old and still has a long way to go before it reaches the same maturity as the electricity networks. We often delude ourselves into thinking that the current protocols and infrastructure are here to stay. Yet, in truth, we need to make many more mistakes before Internet services can achieve the ubiquity and ease of use needed for true consumer acceptance. What's the Internet equivalent of AC power? One thing's for sure: many more years of infrastructure evolution are needed to enable new high-bandwidth consumer services.

It can take many years to establish even the simplest protocols. When the telephone was first invented by Alexander Graham Bell, the technology was so new that nobody knew how to answer a ringing phone. People would stare at it, pick up the receiver, listen for a while, then hang up, frustrated when nothing could be heard. Bell tried to overcome this by training people to call out "Hoy-Hoy" when answering a phone. It sounds incredible, but it wasn't until years later that Edison's marketing machine successfully introduced the word "Hello" instead. (He promoted the use of the word "Hello" by making his telephone operators wear a badge saying: "Hello, my name is...")

Standardizing Internet protocols and infrastructure is a great challenge, but it represents the first step toward true global ubiquity. Technology now progresses so rapidly that it's often outdated before it can become widely accepted. Once the rate of change slows down, it will become possible to create the great power companies of the Internet. Today, if you visit the great data centers of the Internet, you'll see huge bundles of cables disappearing into a forest of cages, filled with equipment from many different suppliers with various degrees of compatibility. Building a service based on today's technology is like a Tinkertoy problem, but the pieces don't necessarily always fit.

The current rate of change has led to the flowering of a lively industry providing the latest gadget to make your network work for you, but the result is a continuous tug-of-war between centralization and decentralization, replication and caching, software and appliance, insourced and outsourced services, thick and thin computing, low and high bandwidth. The challenges involved in building Internet infrastructure have created a great need for tools to manage complex distributed infrastructure systems on a scale never seen before. This trend is likely to continue. It is still early days for the Internet. Many years of change and rapid growth lie ahead of us. But first we need to figure out the Internet equivalent of "Hello."