What are the Internet’s dependencies?

In my last post, I looked at the Internet’s energy use, broadly construed, and contended that we might offload other societal functions onto the Internet.  In this post I’d like to consider in what ways that might be an unwise approach given the dependencies of the Internet.

A couple of months ago, Greer argued that complexity limits might make otherwise sustainable-looking systems falter.  That is, suppose a system uses very little energy, but requires a set of social arrangements, manufacturing and mining systems, and engineering knowledge to make it work at all.  Those complex dependencies might be the system’s undoing even if it doesn’t run out of energy.

The Internet is composed of a complex array of hardware and software assembled around the world with materials, energy, skills, and designs also from a global resource base.  With this complexity comes vulnerability to threats: disruptions in global supply chains, energy shortages, societal instability, and government intervention into its operation.  The last two decades have been relatively placid compared with the previous eight or previous eighty, and have provided an environment in which the Internet has been able to grow at a phenomenal pace.

In the spirit of appropriate technology, a colleague of mine, Shaddi, and I decided to analyze the dependencies of the Internet, down to their roots.  Our goal was to understand where we might intervene in the system to make pieces of it more locally-reproducible, and thus more resilient.  The result was a bit overwhelming, but perhaps a wake-up call.

We looked at a use case of two people communicating over the Internet over a long distance – the dependencies of that scenario are shown in the diagram below (click to zoom).  Each box denotes an abstract or concrete component, resource, or function of the Internet or one of its dependencies.  Arrows denote dependency; dashed arrows denote optional dependency (a “one of these” relationship).  At the top is the use case.  We quickly move through the higher layers which represent mostly functional and abstract pieces that we recognize as pieces of the Internet and on to the lower layers which reside in the realm of hardware.  Once we move beyond hardware manufacturing we enter the realm of chemical compounds and natural resources that are required for many of the relevant manufacturing processes, from making fiber optics to printing circuit boards.  We end with ores or otherwise naturally occurring resources.  We decided not to depict operational or deployment dependencies since they tend to involve a small number of processes like power generation and transportation.

Click to zoom in.

Where the dependencies come from is just as important as what they are – the map below shows the origins of some of the many resources from the dependency diagram above.  The idea of global commodities and international trade are so firmly entrenched in our thinking today that we take for granted that a resource in another country can be tapped if the price or force is right.  However, this is a modern notion, one whose history is scarcely four decades old; higher energy costs and increased resource competition might complicate the picture in the decades to come.

Click to zoom in.


We considered a number of ways of eliminating these dependencies, though I’ll have to go into that another time.  Suffice it to say that we didn’t find particularly elegant ways of replacing the complex and global dependencies the Internet has today.

Leave a Reply


Responses to “What are the Internet’s dependencies?”

  1. Boy, I’m glad to see someone tackling this subject! I’ve been following Greer’s blog for quite some time, and have been sensitive to his general assumption that the expected lifetime of the Internet from now is pretty short.

    I think the track you’re on is a good one, and I’ll be interested to see where you go with it. In addition, though, I’d like to see some good work on another track: if we can’t keep the current, highly complex Internet going for much longer, can we manage a “controlled descent” to a useful form of digital communication with a longer life expectancy?

    I started my software career in the dark ages of digital communications, and have watched the growth of a variety of networking technologies, such as uunet, PC networks like Fidonet, Arpanet, and finally Internet. Thus, I’m pretty confident that there could be a pathway to a low-bandwith, relatively low energy, resilient network that can maintain global communications for quite some time. Whether we can actually follow such a path is another story.

    One of the hopeful signs I’ve seen is the growth of “open source hardware” and shoestring hacker labs putting together a variety of hardware and software. Put that together with ham radio, mesh networks, etc., and it might add up to such a path. The piece that seems to be missing among the techies working on such things is the awareness of the emerging crises, and thus the need to work with constraints on energy and resources to come up with something useful and resilient.

  2. Don – great to hear from you. I’ve been wondering the same questions as you – about what a low-tech Internet might look like – and have been starting to explore it. In the paper we wrote that this dependency analysis comes from we have a very preliminary answer to this question, but at the moment it isn’t very promising. (The short version is that we think a “salvage Internet” that uses existing parts, etc. is very do-able. However, a truly sustainable Internet isn’t likely unless we have some way of replacing IC manufacturing with something less complex.)

    Among my various research plans, I’m hoping to actually try building such a resilient, low-energy alternative Internet in the coming year or so. There are already projects of this flavor for less industrialized nations, but few in the U.S. or with the same objectives. I’ll try to write more about this once it progresses.

  3. The diagram of dependencies is very low resolution and difficult to read even when clicking it. Do you have a version with higher resolution?

  4. Flute — Sorry about that. Looks like it got downgraded in quality somehow. I’ve re-uploaded a high quality version and updated the image / link.

  5. Another point has to do with partial rebuilding of the trunk network. Will we have the effort, will, resources and political ability to lay down new fiber optics and switching as and when existing lines are cut by geotechtonic events or when exisiting country hubs take themselves out of the picture whether due to societal collapse or restrictive government. Avoiding Japan, or Singapore or the Southern Mediterranean could have devastating impact on the existing internet capacity. We witnessed an example in the past when a single trunk was cut in the Mediteranean Sea.

  6. Alice Friedemann March 2nd, 2014 - 2:39 pm

    I’ve been writing since 2006 about how the microprocessor will not survive declining fossil fuel resources and the need we have to try to preserve knowledge now on something more permanent than microfiche or acid free paper as you can see in my article “Peak Resources and the Preservation of Knowledge” at energyskeptic .
    I also have many articles on the dependencies of microchips under the menu item 3) or a fast crash, Microchip fabrication stops