Imagining the Internet Project

  Responses to this 2020 scenario were assembled from Internet stakeholders in the2008 Pew Internet & American Life/Elon University Predictions Survey. Participants were encouraged to provide a written elaboration to explain their answers; they did not always do so, but those who did provided detailed predictive material. Some respondents chose to identify themselves; many did not. We share some—not all—of the responses here.If you would like to participate in the next survey, mail andersj [at] elon dotedu; include information on your expertise.  



Scenario Seven:
The Evolution of the
Architecture of the Internet


Prediction:
Next-generation research will be used to improve the current Internet; it won’t replace it. In 2020, the original Internet architecture is in the continuing process of refinement—it hasn’t been replaced by a completely new system. Research into network innovation, with help from the continued acceleration of technologies used to build, maintain, enhance, and enlarge the system, has yielded many improvements. Search, security, and reliability on the Internet are easier and more refined, but those who want to commit crimes and mischief are still able to cause trouble.

Compiled reactions from the 1,196 respondents:
80% Mostly agreed
6% Mostly disagreed
14% Did not respond

Expert respondents' reactions (N=578)
78% Mostly agreed
6% Mostly disagreed
16% Did not respond

Overview of Respondents' Reactions:
Internet evolution will continue; there will be significant enhancements and updates; a "new" system will not "replace" the current architecture; transformations will occur gradually. IPv6 and the Semantic Web will be part of this continuing development of the system. By 2020 more-specific "walled gardens" may be created to maintain network control and for various other reasons; there may be split networks or partitions. Governments and corporations are expected to leverage security fears to retain power over individuals. Every system generates unintended consequences—the more you have to gain, the more you have to lose—crime, piracy, terror, and other negatives will always be elements in an open system.

Below are select responses from survey participants who preferred to remain anonymous. This is not the full extent of responses. To see more, read the report PDF, and to read reactions from participants who took credit for their answers, please click here.

Current Internet will solve its own problems through evolution. Governments, by definition, are always second-generation and to that degree self-defeating. Security is a trade-off. Having anti-virus, anti-spyware, firewalled computers is a trade-off for living in a Microsoft world. Trade some of your touchy-feeling conveniences for a more robust (and community, not government evolved) system like Linux/BSD and you have more robust security.

There will be two Internets—one for government, libraries, education, where you have to have a valid username linked to your "real-life" identity, and a second, Wild-West Internet for the viruses, phishing, pr0n, recreation. [pr0n is slang for porn, originated when spam filters began searching and destroying messages with "porn" in them]

Internet2 will always be academic. Actual Internet and its technologies (TCP/IP) will still be with us for a long, long time.

100% totally agree. There's way too much investment needed to do an overhaul, so next-gen Internet will be more of an overlay. I could see the key backbones being upgraded, but all of the other “pipes” require so much more. WiMax (wireless broadband) infrastructure will increase and help, but that is also an “overlay” technology, not new infrastructure.

It's hard to imagine how a next-generation Internet will eliminate viruses, spam, phishing, and worms.

The Internet is now too large and ubiquitous for a "flag day." The results of "next-generation" or "clean-slate" projects are therefore likely to have their primary impact on the current Internet (IPv4 + IPv6) than on a greenfield replacement for it.

GENI's purpose is certainly to improve communications—its goal is not to replace the Internet (per se). If that happens, great, but it isn't GENI's purpose.

Even if a brand-new system was invented, it wouldn’t be a substitute for the current one by 2020.

The roads may be improved, but they'll follow the same routes.

This is splitting hairs. We'll see a lot of spectrum open up and governments will begin using those in different ways. Will this be "one Internet"? What counts as "the Internet"? Arguably, there are many Internets right now. Regardless, a lot of social divisions result from people not experiencing the same Internet. Still, things will improve on all fronts with regard to security and reliability and search.

I see a two- or three-tiered system evolving.

Virtual worlds and 3-D presence will create another layer of identity and trust to our online interactions.

GENI is not unlike the DARTNET, CAIRN, and PlanetLab of the past. They will be used to prove out ideas, but will not become the fundamental basis. That said, yes, there will have been great improvements in Internet service.

The Internet has from the beginning grown by evolving rather than through revolutionary change and I don't see this changing. Previous attempts to "replace" the Internet have failed miserably.

I doubt that anyone would support having to move to an entirely new system of Internet just because of the huge hassle it would cause during the process of switching between the two.

Thirteen years is not enough time to replace the Internet and access networks. IPv6 will perhaps be well-deployed and we will have new technology from this R&D to make the Internet more secure etc. But a completely new infrastructure will be unlikely.

The Internet is too big and is too decentralized to be upgraded to a new system. We'll be lucky if we're at IPv6 by then.

The complexity and dynamism of the current Internet makes it beyond centralised top-down control or straightforward replacement.

Absolutely, it's going to be an incremental and evolutionary process.

The ones causing the troubles are going to refine their efforts at computer crime, making it all the worse.

The biggest error here is thinking NSF is doing anything. CHINA! Or perhaps a China vs. USA virtual war, but then it wouldn't be NSF it would be DOD running the Internet, which isn't that different than how it got started. The Internet will always be hackable and open to crime. It is naive/silly to think otherwise. That it might be the battleground between countries or insurgencies, is totally believable. So not just individual criminals but collectives are acting as criminals.

A completely new architecture will be built, but it will be more controlled and the current Internet will be the underground.

A global transition to a completely different Internet architecture will not happen in the foreseeable future.

I suspect that there will be investment in dedicated safe networks driven by commercial demand, but the cost will delay introduction. And the benefits are key. Incidentally, I am reminded of one of Parkinson's laws: When [the Internet] is perfect for [what people use it for], it's dead (i.e. people are using something else, innovation and change are no longer happening). So almost in introduction, the next-generation, perfect Internet will have to cope with things it was not designed to do.

New research will yield better designs but not better by enough to mobilize sufficiently enough to permit these new designs to get much beyond the labs before some patch to the existing system sucks the wind out of them. There is too much invested in the current infrastructure to allow any change that is not hugely beneficial.

There will be technologies to augment the existing Internet not replace it.

Well, yeah. After all, you can still plug a 1930 phone into the current network and it'll still work.

There will be new ways to prevent and detect online crimes and mischief.

There will always be bad people as well as government officials wanting too much control.

The analogy I would use is the human body. We replace our parts in a continuous stream, with cells dying and new ones taking their place. The management of the Internet will function similarly—we will never flip “the switch” to a completely new Internet, but rather will keep changing out parts as technology advances.

History teaches that installed processes and investments are terrifically difficult to replace in the absence of radical changes in functionality. People are going to take the Internet of the PC in order to gain some more security? I don't think so. Incrementalism will win out.

Due to costs and potential civil liberties challenges, the next-generation Internet architecture will likely not appear by 2020.

Couldn't agree more. We now have a billion embedded users.

It is what it is.

Open system a must—this is a major threat to governments in general, and especially communist and controlled governments. I see a huge effort to develop a new network with much more controls—and in this case the horse is out of the barn and it will be very difficult and expensive for any alliance to change the basic structure.

The disruption from replacement would be catastrophic.

2020 is really not that far away in terms of revolutionizing the architecture of the Internet. Evolution is more likely.

The current Internet was not designed to fulfill the needs of the future. We need a new one.

Developing a second Internet would cost too much time and money, and would ultimately fall victim to the same problems the current one does.

Those who want to cause trouble will always find a way to do so.

It won't change all that fast.

Those who want to commit crimes and mischief are ALWAYS going to be able to cause trouble if they try hard enough, no matter if it's a new system or an old system. New systems simply have new bugs to be discovered and worked around.

It will change for sure. But will it still be the Internet? Could be something else entirely. After all, in 1982 nobody could have predicted the Internet at all!

It is hard to see a complete replacement of a system in which there has been significant investment; particularly with the current economy. Current investments in cyberinfrastructure do not seem to be going that way in the US.

Social tagging and other social research applications will play a huge role here.

Crime has always been a feature of human life and will most likely continue in some form or another.

Though 95% of the population will use the Internet for good, the 5% who choose to use it for evil will make security a continued issue. Those who have not been able to connect meaningfully with others may also find themselves so disenfranchised that they behave in Columbine kinds of ways.

I don't think we can even imagine what the format will be 12 years from now. I did not think in 1982 that I would ever be able to afford a computer, yet within thee years I was paying for one on-time.

Intelligent crime will increase with technological advancements.

Present IPv4 will be replaced by IPv6 by 2011. After that, research into a new network will start, and by 2018 a new network will be started up.

This sounds like the same chance as me having a flying DeLorean.

It really is beyond government control—or at least beyond any particular government to reform.

It's human nature…someone always tries to cause mischief in most systems, from the errant school kid to the sophisticated hacker.

Consider what the Internet was only 5 years ago.

The Internet is such a part of life in the world now, I'm not sure it can be replaced. It is tied to every sector and every facet of life. People will not easily move to a new system that they will have to relearn and that would be more restrictive.

It will depend if Google wants to do it!

Too much can be achieved by replacing the existing Net not to have built the future Net.

Historically we tend to improve what works for infrastructure rather than replace—and the Internet works just fine.

Very large systems are difficult to replace due to the resources required. We will continue to improve the current technology, not replace it. However, over time, the improvements—due to their complexity—will appear as though the entire system had been replaced.

There will always be miscreants, but efforts to improve the current Internet are ongoing. (Highly irritating to think of all the resources wasted on combating spam!) Can't see how a next-gen Internet could completely replace the current one—too many entities would have to agree and cooperate!

Replacing the Internet is impractical. The power of the network is its diversity. The new tools will just have to integrate to the current network.

The technology can be developed and implemented incrementally. I would love for governments and industry and individuals to cooperate and make a leap, but that's a lot to ask for in 12 years. (Translation: the technology may be there, but the “powers that be” will have difficulty cooperating and power sharing.)

Online security will always be a concern. If a "next-generation" Internet is developed, it may be used for some types of interaction, but people will still prefer the original.

I can't see that the funds would be available for this by 2020.

The Internet as we know it will evolve more than it will be replaced. Fiber optics may replace copper. TCP/IP may be replaced by something far more efficient. The GUI layer will change. But, I don't think we're all going to go to bed one night and wake up the next morning with a new Internet.

A complete overhaul/change in the system would require more funding and disruption than the public could handle.

Enhancements are good as long as they don't infringe on civil liberties. There really is not a way to completely squelch mischief, thankfully.

The Internet of today is nothing like the Internet of 10 years ago. In 10 more years, I can't imagine that it would be the same as now. The urge to constantly be faster, do better, etc., will always be there.

Of course it will. The Internet has become the primary source of research in the last 10 to 15 years. Though that still does not guarantee substance of content or its validity.

I think it possible that something new will evolve.

There could be multiple systems.

Technological evolution will follow the model of biological evolution, building more complex structures on top of the simpler base without eliminating it.

Because the original Internet infrastructure only reached the remotest areas in 2018, it will be a while before the NGI (next-generation-Internet) replaces it.

For a minority, a new system may be being used, but I think most people are still trying to come to grips with the current system, and will not be open to a new system by 2020.

It will evolve, not replace. We have come a long way but more research is needed.

There will always be a need for someone to build to better mousetrap.

I hope everyone will fight against the build of a whole new network with whole new rules.

First they have to agree on what to do.

If you build a better mousetrap, humans will still find a way to outsmart it.

"Those who want to commit crimes and mischief" will ALWAYS be able to cause trouble. And most people would prefer the existing infrastructure to one where "the trains run on time."

The Internet routes around censorship; it also routes around security. People are lazy and gullible. That won't have changed by 2020, so security threats will still exist.

It is too difficult to put the genie back in the bottle.

New technologies, new hacks. A totally new Internet will require new technological standards, and it takes time.

Plus ca change, plus la meme chose. Hopefully, we will have better controls in place to combat issues like spam and viruses.  But their will always be evil ones seeking to exploit weaknesses.

Participation culture—like Google, for example—will create a seamless interface for ongoing development.

A second secure network will surely be developed as a response to hackers.

It is human nature to attempt to disrupt systems such as the Internet. If humans can invent it, then humans can disrupt it.

Yes, the "last mile" may be WiMax or fiber, but the essential structure will remain.

I see the current Internet following Milton Friedman, libertarian economics, and eventually we will be able to effectively police the community and create a world where there is more opportunity through the legitimate channels. As Internet giants become bigger and bigger they will stay on top of research trends, and any innovative technologies will immediately be introduced through our current Internet.

There is a move afoot to examine popularly held distinctions between crime and creative rights.

Since there is so much dependency on the Internet (phones, medical awareness, communication with others, conducting business) having a new Internet seems unlikely.

Just like the health care system that is so engrained in our society, the fundamental architecture of the Internet will not be changed. There would be too much push back from a great number of individuals, not just those concerned about civil liberties.

Unfortunate scenario - but probably accurate.

The idea of a “clean” Web is a long way off. Although there are certain aspects of the Internet that almost everyone would like to see disappear I doubt that it will happen soon.

Most of the Internet infrastructure is the same infrastructure we have been relying upon for ages. It will take a very long time before there will be some massive change to the current architecture—largely because it is not only the US and Europe that are online. OK?

Most clean-slate design efforts are fundamentally flawed in that the present set of problems of the Internet are the other side of the very strengths of the Internet, openness, intelligence on the edge, flexibility, low-cost, etc. You can have all both ways. There are problems with the computer host environment as well. Thus, addressing the problem at the right place and striking a good balance of the up and down sides of the same important character of the Internet should be the focus. The Internet is no different from any other way to communicate and share information.

The last thing anyone should wish for is government regulation of the Internet. The government has time and again demonstrated its total incompetence in the construction and management of technology; we should want to keep them as far from it as possible.
With the work in virtual firewalls advancing, I don't see the need to create a whole new Internet. It's always been a dynamic entity.

Spy vs. Spy.

Governments will build the next Internet just as they built the one we have and, this time, having had a glimpse of what it can do, they will look to develop it to their advantage.

I doubt the world's governments and commercial entities will ever be sufficiently coordinated to enable the generation of a new Internet. The changes are for more likely to be patchworks as entities attempt to fit acquisitions into their existing structures.

Create a new Internet? The current one is moving too quickly and I don't foresee anyone catching up and creating something people would use, not to mention the fact that this would require international governments working together. More likely the current Internet will continue to exist as some sort of cool underground where you can do what you want.

2020 isn't that far off. Things don't change that quickly. 1996 is as far off as 2020, and the network isn't substantially different. Base technologies are the same, as are taxonomies and infrastructure.

No single government or corporation will be able to finance requirements to revamp entire Web. As such, progress will be made piecemeal and by refining the existing Web. At same time, as long as incentives exist for mischief (e.g. financial, intellectual, power/control, others) then troublemakers will continue to act.

As long as there is little “adult supervision” on the Internet, it will continue to be bogged down in spam. There must be strong oversight. It would be much better if users were involved in providing this, instead of acting like spoiled toddlers and insisting on libertarian notions that have clearly not worked so far. Also, as long as big business is allowed to be unregulated on the Internet, it will continue to suck up all the bandwidth it can, leaving less and less for individuals.

There is no fence tall enough for a man who wants to cross it. It will be too costly and impractical to build a new Internet from scratch since it is impossible to ignore the importance of the current Internet, as imperfect as it is.

Physically, the Internet will continue to be a network of networks. However, there will be some safe tunnels for corporations and other groups through the current networks that provide better speed and assured authentication. These safe zones will blend seamlessly with the rest of the Internet. New forms of viruses and security breaches will be problematic. Law enforcement will have special units to police online crime.