A Plumbers View of the Future of the Internet

©Quinn Stewart, 11/3/2000

A number of years ago as a young boy, I was invited to a presentation by Bell Laboratories, which was then touting the future of a new technology it was working on called fiber optics. It was the early '70s, and the demo included some fiber optic "sculptures" of trees and whatnot. It was quite expensive at the time, but looked cool.

Life intervened, and I went on to become a plumber. I started out working in the oil fields of East Texas (I promise this story won't go on forever, and it WILL make some sense) working on high-pressure natural gas wells, and the pipelines that connect them. This was the same area where my grandfather began his career as a pipeliner in the 1920's. When the domestic energy economy tanked, I came back to Austin and worked installing underground utilites, and later in commercial and residential plumbing.

Pipes and pipelines all must obey the laws of physics. If you need more oil, water, or gas, you either need a bigger pipeline or have to increase the pressure on the line to push more through it, or both. This leads to greater expense.

Electricity is roughly analogous to plumbing. If you need more electricity, you need bigger wires or higher current to increase the capacity. It is a little different in that it is often cheaper to use a higher voltage, but suffice it to say that it is a combination of the size of the "pipe" and the "pressure" inside of it that dictates the amount of oil, water, gas, or electricity you can push through it.

You also have to route this stuff. Pipeline stations, electrical power grids, water and gas distribution systems all involve valves and switches that must match the capacity of the lines feeding them, and route the resource to its proper destination. The larger the capacity of the system, the more it costs to deal the laws of physics, which dictate that friction and resistance will increase costs proportionate to the capacity of the system.

Enter fiber optics. Telephone communications since the beginning have been carried on copper wires, and still are today. The Bell Labs research led to the use of pulses of laser light traveling thru fiber optic cables to carry telephone conversations. The first telephone transmissions using fiber optics occured in April and May of 1977, by GTE and Bell Telepone. As laser technology evolved, and manufacturing techniques for fiber optic cable improved, the costs for fiber optic transmission of telephone and television signals dropped. By 1985, single-mode optical fiber was being widely used for carrying voice traffic here in the US.

So what? Well, fiber optic cable uses pulses of laser light to transmit data. The laws of physics begin to get a little muddled where light is concerned, so it is a lot more effecient to use optical transmission than electrical transmission. Huge, thick bundles of copper cable were being replaced by small bundles of fiber optic cable, able to carry the rapidly increasing amounts of voice and television traffic. Parallel to the development of fiber optic cable for telephone transmission was the evolving use of the medium to carry data traffic between computers. The rapid expansion of the Internet would have been impossible without the growing use of fiber optic cable.

What does all this have to do with plumbing, or the future for that matter? Plenty. Fiber optic cables are the information pipelines of the Internet, capable of carrying voice, video, and data traffic simultaneously at the speed of light. Enron Corporation has operated natural gas pipelines for years, and is now using their extensive rights-of-way for a fiber optic network as well. But there are some interesting differences between gas and information pipelines. If Enron wanted to move more gas thru their pipeline, they would either have to make it larger, or operate the line at a higher pressure, or both. This is not only expensive, it can be time consuming and dangerous. The physical properties of light are quite different from natural gas, or water and electricity for that matter. This makes optical fiber an interesting resource. For Enron to increase the capacity of its fiber optic network, it could add more optical fibers and more lasers. Wireless and satellite technologies hold some promise for transmitting large quantities of date, but nothing even close to the capacity of fiber optics. Because we are dealing with light here, there are also other possibilities.

The Future of the Internet

A large portion of the optical fiber network built worldwide from 1985 onward remains in use today. The seemingly insatiable desire for bandwidth has led many companies to expand and build fiber optic networks. There are at least two factors, both related to the unique physical properties of light, that lead me to believe that bandwidth capacities will increase rapidly for the foreseeable future.

One of these is Dense Wavelength Division Multiplexing, or DWDM. Much like a prism in the sunlight divides light into its various wavelenghts, DWDM uses combinations of lasers and other optical means to divide pulses of light into a multitude of wavelengths that share the same optical fiber. Imagine a pipeline that used to carry only water, that can now carry 211 flavors of water, milk, orange juice, etc. all in the same pipeline. This is the basic promise of DWDM. Both Lucent (formerly Bell Labs) and Fujitsu have demonstrated multi-terabit DWDM systems that should lead to commercial implementation in the near future. Multi-terabit capacity is nearly unfathomable, imagine the current entire world-wide Internet traffic per second traveling over a single fiber optic cable, and you are just using one-third the capacity of a single terabit DWDM fiber.

The other factor involves routing. For many years optical fiber terminated in electrical switches, which greatly slows down the data transmission rate. Information traveling in fiber optic cable must be converted from photons to electrons in order for switching to occur. A single or multimode fiber optic cable can include many of these electrical switches and routers across the network. The use of optical switching could greatly increase the capacity of existing fiber optic networks, and is a necessity where DWDM is concerned. Current research by Lucent has produce optical switches using tiny mirrors, but a need for optical switching on DWDM networks is needed. Research at MIT has led to the development of photonic band-gap crystals, which show great promise in making optical switching of photons a reality. This would allow our pipeline full of 211 flavors to be effeciently routed to 211 different destinations.

The potential combination of these two technologies holds great promise for the technological future of the Internet. Continued expansion of network capacities combined with ever more powerful computers should result in Internet traffic that dwarfs that of the previous decade. Limitations of the current Internet will rapidly dissappear, and in their place will be....What?

What will be the Future of the Internet?

One effort to answer this question is underway in research being done here at UT and elsewhere on Internet2. Internet2 is meant to be a testbed for developing advanced applications for research and education among its member institutions. The primary Internet2 network is called Abilene, and it is currently implemented here at UT at the router level. Internet traffic here at UT that is destined for another Internet2 participant is routed over Abilene rather than the commercial Internet. Abilene is currently transitioning from an OC-48c backbone to a OC-192c DWDM backbone, which will operate at 10 Gbps. What does THAT mean? It means that the network capacity of current research efforts is minisicule in comparison to what has been demonstrated experimentally. In other words, by all indications, network capacity will not be a permanent limiting factor in the future of the Internet.

The Real Question about the Future of the Internet

What are we going to do with it? Imagine an oil pipeline like my grandfather used to work on increasing in capacity by a factor of 211 times in less than a decade. Never gonna happen. Yet the Internet and the trickle of information it contains today is doing just that. Our role as information professionals will be to work towards utilizing this unique resource in ways that benefit society, just like we have always tried to do. My goal here was hopefully to give you some background and insight into the technological future of the Internet. As a student at GSLIS in early 1998, one of my projects concerned the implementation of cable modem and DSL service here in Austin, which was met with great skepticism. Today these technologies are commonplace. The pioneering use of streaming media for undergraduate education here at UT is but one attempt to better utilize the current network capacity we have, and learning about its effectiveness and appropriate use in the process. While I by no means claim to have an accurate crystal ball, I do claim that any plumbing system that both greatly increases its capacity and lowers the cost of construction and maintainance will return benefits to its users. It is up to us to learn how best to use it, and to help others do the same.