Creating a Backboneless Internet?

Peter Trepan asks: "The Internet is the best thing to happen to the free exchange of ideas since... well... maybe ever. But it can also be used as a tool for media control and universal surveillance, perhaps turning that benefit into a liability. Imagine, for instance, if Senator McCarthy had been able to steam open every letter in the United States. In the age of ubiquitous e-mail and filtering software, budding McCarthys are able and willing to do so. I Am Not A Network Professional, but it seems like all this potential for abuse depends upon bottlenecks at the level of ISPs and backbone providers. Is it possible to create an internet that relies instead on peer-to-peer connectivity? How would the hardware work? How would the information be passed? What would be the incentive for average people to buy into it if it meant they'd have to host someone else's packets on their hard drive? In short, what would have to be done to ensure that at least one internet remains completely free, anonymous, and democratized?"

Comments

You're on it baby..

[brokenin2]

It would look an awful lot like the internet we have now.

You're describing the original design of the internet, which we're still running with essentially.

In practice though, it would be insane to let everyone with a DSL line to two different locations update routing table through the entire internet. The mechanisms to allow this exist (bgp, ospf) but major ISPs that don't want their network to fall apart prevent it because their service would quickly turn to crap. ISPs with missing filters have actually caused internet wide splits, when the entire internet tried to route through someone's T1's connected to two different ISP. BGP with a little better cost system could help that, but anyone could still cause a split anytime they liked. Think of an entire internet that acts more like IRC.

The core of the internet is still just a bunch of peers, but if you want things to stay up, they've got to be a select group that really know what they're doing. You're still free to peer directly with anyone you want, just don't expect everyone else to use your internet connection to get there too. Most people don't want to have to buy two internet connections for marginal gains anyway.

Perhaps a software solution like TOR or Freenet could help you sleep better at night?

Re:You're on it baby..

[ZagNuts]

Perhaps a software solution like TOR or Freenet could help you sleep better at night?

Don't know much about TOR but I just thought I'd clarify about Freenet. It is indeed a software solution to what you are asking about in which the sites are accessed in an entirely peer to peer manner. Instead of having static routing tables located at specific points each computer in the network maintains its own routing information. If a computer doesn't know how to get to a certain site it guesses by asking a neighbor if it has the desired data. Data is cached throughout the network so that sites are stored as distributed files, meaning at any one time if your computer is a part of Freenet it could have information related to a number of sites.

The good thing about Freenet is that site accesses are entirely anonymous. There is no way to be traced AFAIK. One of the bad things is that it takes a computer a long time to build up enough routing information to access any websites at all. You have to run the Freenet program for a few days before you are able to access anything and even though its painfully slow. The other problem that people have is that you have to store any content that goes through your computer. Freenet is plagued with child porn sites because the anonyminity that it provides. This means that if you are running the freenet program you are likely to have child pornography data stored on your computer even if you have never visited those sites. While the legality of this is questionable, the ethical issues are obvious.

Still it is a very interesting concept and definitely has its applications (China anyone?).

Re:You're on it baby..

[jovetoo]

I agree... let the service providers provide the service. If you want privacy, use encryption. Unless some higly specialised entities have developed quantum computers and kept it a secret, they won't be able to break it in any time frame suitable for mass communication snooping.

Re:You're on it baby..

[Mattcelt]

I think that response may have missed the point of the submitter's original question. I read it as "is there a way to prevent all traffic from traversing predictable routes and hubs, thereby disallowing any entity from collecting all of one's transmitted data and using it against one?"

Essentially what the submitter is interested in is a meshed network, which to my knowledge is the only network topology yet created which does not use hubs, centers, or buses to carry conglomerated traffic. Remember that things like bittorrent, bgp (less so), and other similar protocols are really creating "virtual" meshes, not real ones - all of your traffic (and that of every other person in your segment) is still travelling to your ISP, and that to their backbone. So anyone who sits at those hubs or backbones would be able to see all your torrent traffic, and who it is going to/from - it is only the separation of the ISPs and the RIAA/MPAA/FBI that keeps them from knowing your every move on the Internet! (Encryption and proxies help, but it aren't a foolproof solution, btw.)

Also, TCP is designed to be fault-tolerant, but also semi-optimizing, taking the shortest perceived route to its destination. So unless a backbone is down, most (if not all) traffic from you to a host between which the backbone sits will travel on that backbone, very predictably. TCP is not privacy-sensitive.

The short answer is that in a wired world, there is no feasible way to create a mesh. The strength of the mesh is algorithmically tied to the number of other nodes each node is connected to. So unless you're going to dig up the yard between you and, say, three of your neighbors, and they and two more of theirs, and so on, across the entire country, you will end up with a topology which looks more like what you've already got, with a smaller number of larger rings and stars, each funneling through a central location.

In a wireless environment, the possibilities are much better. Some police precincts in the U.S. have been experimenting with mesh-networked radios, where each radio is a repeater as well as a transceiver. Thus a linear configuration of radios could extend the range from perhaps a 30-mile radius to a 60-mile-per-radio diameter for as long as the chain is unbroken. This isn't the optimum configuration, however, since it is presumed that one would want redundancy, so you would be forced to configure the mesh in such a way that you could talk to at least three other nodes at any given time. This requires a very high density of nodes, so it would work much better in a more densely-populated area than one nodes are scarce.

I hope that answers the question.

Re:You're on it baby..

[r_naked]

In practice though, it would be insane to let everyone with a DSL line to two different locations update routing table through the entire internet.

We seem to be scaling rather nicely.

http://anonetnfo.brinkster.net/

Re:You're on it baby..

[mrchaotica]

The only trouble is, it's not child porn until a court of law rules it as such. Therefore, at the time you'd be deciding to filter it it isn't illegal yet, and that means you're just filtering things you don't agree with arbitrarily. And if you're doing that, what's to stop you from filtering other things you don't agree with, like websites advocating equal rights for minorities (hypothetically, that is -- I'm not calling you a racist or anything!)?

In other words, if you're a common carrier you can't make any decisions about blocking (allegedly) illegal content at all, because it would be too easily abused.

Re:You're on it baby..

[Alsee]

Perhaps a software solution like TOR or Freenet could help you sleep better at night?

Nope.
Are you familiar with Trusted Network Connect?

It is a new specification from the Trusted Computing Group to control and restrict network connections, and to control and restrict the networked computer.

"The TNC architecture enables network operators to enforce policies regarding endpoint integrity at or after network connection."

Of cource the Trusted Computing Group is advertizing it as a good thing, and is advertizing it as prortecting against viruses and network attacks, etc. However it is an incredibly powerful system to impose general restrictions and controls. Aside from being able to impose a global DRM system, it has the power to restrict and control and ultimately defeat TOR and Freenet and any other networked program you care to name.

Microsoft has already issed a press release that they are implementing this system.

The US President's Cyber Cecurity advisor gave the keynote speech at the Washington D.C. Global Tech Summit and the main thrust of his speech was to call on ISP's to plan on implenting exactly this sort of system. He called on them to implement such a system to fight viruses and to secure the "National Information Infrastructure" against Terrorist Attack. He called on them to make it a mandatory part of the Terms Of Service for internet acces. And the Global Tech Summit audience applauded his speech.

The EU and the UN have been running a large number of international workgroups on DRM and on establishing a new "Information Society". An Information Society which is to include exactly this sort of network control and DRM enforcement system. EU and UN have been running many workgroups on to work out a new system of Internet Governance to set up and manage this new Information Society. And in case you hadn't noticed, the EU and UN have been pushing pretty hard lately to remove control of the internet from the US and to place that control in the hands of a new UN Internet Governance organisation.

Intel, AMD, and IBM are all building new CPUs with this new Trusted Computing control and enforcment system built in. And it appears that by the end of *THIS YEAR* that all new new PCs will come standard with have this Trusted Comptuing DRM enforcement chip welded to the motherboard, if not built into the CPU itself. The hardware specification for Windows Vista requires this encorcement chip on the motherboard for full and correct Windows operation. And no PC manufacturer and no PC retailer can possibly survive selling new PCs that are not Certified Windows Compatible and which do cannot properly run the latest version of Windows. They cannot realistically survive selling hardware where Windows spits out error messages stating that that you have incompatible hardware, error messages saying that the full featured graphics interface and thenew hires graphics do not work because you have incompatible hardware.

Five,seven, ten years down the road the internet absolutely can be developed in a direction to defeat TOR and Freenet. And there are several hundred powerful coroporation, and many governments and international organistations that see that as a GOOD direction to go, and which are actively and forcefully pushing to establish such a network.

And the way to establish such a network would be to establish an international body for Internet Governance (the world would obviously never accept such a system imposed by the US), and for that international standards body to establish international agreement on new internet standards similar to or including Trusted Network Connect, and to establish such a system along the internet backbones, and from there to push it to the ISPs, and from there to have ISPs impose Trusted Network Connect on all connections. It would then be impossible to connect to the internet unless you are using the mandatory enforcment hardware and software

Bad Idea

[Kasracer]

If Bit Torrent is of any example, this would be a bad idea. One day you may be able to get to Google fast and then the next, it may take forever to load.

Peer to Peer internet would be horrible. Not only would it be unreliable, but at time slow.

Sure some agencies can access our information because it's centralized, but if we don't want them to see something, it's not hard to encrypt it. Hell I'm even working on an encryption application.

Re:Bad Idea

[Firehed]

But encryption is a waste of time if you can bypass the evildoers entirely.

The main problem with a P2P internet would be bandwidth, at least at this point. There just aren't the resources available - hardware or software - for people to be running /. out of their mom's basement. Even a good amount of small businesses wouldn't be covered by a fairly decent dedicated server, but they can't afford to set up a cluster to run things like a hosting company can, let alone hire someone to set the thing up (or be expected to know how to set it up and maintain it themselves), even if everyone had a petabit connection to the internet for a buck a month.

So it's really not a feasable idea. If it were realistic, it'd be great, but without some major hardware changes (large amounts of solid-state flash-volitility RAM-lifecycle storage at affordable prices) and obviously a complete structual revamp, it couldn't work.

So until we all have streaming-ten-next-gen-uncompressed-high-def-movie s-all-at-once internet connections with equally fast storage that has nanosecond seek times, it's just not realistic. So until that time comes, keep encrypting, and then encrypt over that (because if it's a bitch to crack the first layer of 512, you're screwed trying to break through that second layer of 2048).

Or just lobby for us regaining our privacy. Too bad there are so many people willing to lose every bit of their privacy if it helps to reduce the already-miniscule chance of them being injured in some sort of "preventable" terrorist attack.

Internets!!!

[NorthwestWolf]

More than one internet? Looks like George W. would finally have his Internets!

Not exactly practical

[georgewilliamherbert]

If you need something like a terabit of bandwidth between the US east and west coasts, consider how many peer to peer link chains across the country will be saturated carrying it.

One of the major problems right now in the commercial ISP backbone environment is what happens if there's an outage; what's called route flapping, where routes dissapear and reappear, and all the routers affected have to recalculate how to get to various endpoints, can already saturate the router CPU logic for big, industrial grade room-full-of-racksize-router backbone facilities. Going to a more diffuse network at high bandwidth requirements exponentially makes this worse.

P2P across a city? Not ridiculous.

P2P across the world? Baaad idea.

Circa 1982

[sphealey]

> Is it possible to create an internet that relies
> instead on peer-to-peer connectivity?

You have just describe the net (later the Net, still later the Internet) circa 1982. You can search Usenet to read about the excitement level when USR 2400 baud modems were released: doubling of connection speed to transmit netnews!

Of course, you can also read about what happened when news (alone) was distributed on a meshed basis.

sPh

Yes, but not really.

[khasim]

Is it possible to create an internet that relies instead on peer-to-peer connectivity?

From a hardware/connection standpoint, every single user would have to have a router that could connect, somehow, to every other user/router.

That is the "backbone" and where the "bottleneck" is.

Tier 1s?

[ObsessiveMathsFreak]

It would look an awful lot like the internet we have now.

Except for, you know, the Tier 1 ISPs, on whose networks practically all our traffic passes at some point.

Control them, and you control the net.

Re:Tier 1s?

[brokenin2]

I'm not saying that there isn't a core to the internet. It's there, but that's not by design, it's a convention to keep the internet from totally sucking.

His question was, "Is there a way". The answer is yes, but you don't want it, so people stopped doing it. Anyone can peer with anyone else, but the copper/fiber cost to take the core out of the picture prevents anyone from wanting to do it. If you're worried about big brother, encrypt.

If he really wants what he's asking for, he can start finding peers on the other side of the net, and he can keep *his* traffic off the backbones once he has enough peers (and he's built some enormous route tables as well).

Re:Tier 1s?

[toddbu]

The answer is yes, but you don't want it, so people stopped doing it.

Then what do you make of the Seattle Internet Exchange?

Re:Tier 1s?

[dnoyeb]

I disagree. There will be as many "Tier 1" ISPs as the people need. The only reason there are a few now is because we only require a few. This is all besides the point.

When congress starts legislating your network architecture is meaningless. If your worried about invasion of privacy you should address it with your vote as well as your intelligence. If you can explain the issue perhaps you will get more votes. Its tough to fight the force of the media, but its not impossible.

Re:Tier 1s?

[cat6509]

>>It would look an awful lot like the internet we have now.
>Except for, you know, the Tier 1 ISPs, on whose networks practically all our >traffic passes at some point.
>Control them, and you control the net.

Keep the backbone, without huge aggregate networks the internet is not cost effective and not to mention what kind of routing problems and bloated BGP tables we would have, just do VPN to peers you trust, that can be either router-to-router ( GRE IPSEC hacked-together-ssh whatever ) or somehting even browser based , but fragmenting things into many many more smaller peers just makes things unusable.

Re:Tier 1s?

[Yartrebo]

Tor is unlikely to go mainstream so long as P2P is kept off of it. My guess is that mp3s would create the largest demand for a highly secure network due to the risk of copyright infringement lawsuits. Mp3s are also small enough to accept a 3-1 or so overhead, but big enough to put a heavy load on the system.

Porn (particularly highly illegal types) would also be a strong demand driver. Anything else that a substantial amount of people want and is sufficiently prosecuted or sufficiently taboo will drive demand.

These tools aren't answers in and of themselves though, as they themselves can be banned and filtered out, even if the contents cannot be looked at. Also, monkey-in-the-middle attacks can work if enough nodes are controlled, which major ISPs/government agencies can do.

Tor + a decentralized network would be far more resiliant to attempts to flat out ban Tor.

Re:Tier 1s?

[Schraegstrichpunkt]

Anyone can peer with anyone else, but the copper/fiber cost to take the core out of the picture prevents anyone from wanting to do it.

Not only that, but there's the problem of growing routing tables. Every host on the Internet needs to know what direction to send packets destined for every other host. IPv6 is designed to alleviate some of the problems of big routing tables, but that's just because it makes it easier to map a hierarchical network topology into a hierarchical address space (thus reducing the need for explicit routing table entries).

As far as I know, a global flat address space is practically impossible.

All mail was read in WWII

[Derling Whirvish]

Imagine, for instance, if Senator McCarthy had been able to steam open every letter in the United States.

Before and during WWII all mail crossing an international border in or out of the US was steamed open and read. This included all mail, all packages, all telegrams, and all telephone calls. In addition to all mail being steamed open and read, it was censored if the Army deemed it to be necessary to support the goals of the Army. Letters would arrive with portions cut out by scissors. They also censored all international media -- radio, newspapers, and magazines both incoming and outgoing.

It's quite easy to imagine as it's already been done.

Re:All mail was read in WWII

[Anonymous Coward]

"this included all mail, all packages, all telegrams, and all telephone calls."

The capacity to read everything did not exist.
This was during all out war not some informal war with no timetable.
This data was not kept indefinitely.

Lastly the computing power did not exist for a politician to do an SQL query on your life history to determine if you are "desirable".

Dangerous and misleading analogy.

The Solution Is Crypto

[blofeld42]

Encrypt your email traffic, so that even if it is intercepted it can't be read.

The government can still do some traffic analysis (they sniff headers rather than read the contents of the messages) and they can learn a lot from that, but such is life.

Re:The Solution Is Crypto

[TheBeginner]

As a related matter, I've found myself wondering why encrypted email has not become far more popular - or encrypted IM for that matter. I downloaded and installed PGPMail myself a few years back, but could never get any of my friends to install it as well. This strikes me as strange considering that I know that were I given a choice between an email client with encryption and without, I would choose the former. I assume most people would. So why hasn't this been offered as an automatic part of Outlook or Thunderbird? Why haven't market pressures led to this? Is it technical difficulties? Or is it something else?

Re:The Solution Is Crypto

[friedmud]

As I've said before... some people dont have anything to hide....

If people want to read all the little love letters I send my wife all day... or the email to my Dad about the cool car I saw on the way in to school this morning.... then go right ahead...

What I'm wondering is why people feel the need to hide their e-mail activities. The only situations I can think of are when you need to send sensitive information quickly (the secretary for my advisor asked for my Full Name, SSN, Address and Telephone number through email recently.... I promptly walked up to her office and told her what they were... but people not paying attention _might_ just hit reply)... but people should be aware of those situations and just avoid doing it (or use encryption on those case by case basis).

Think about it this way... when you send something using the US Postal service you can't guarantee that the message won't be read by dozens of people along the way. How many people do you know of that use secret code languages to communicate with regular mail? That's what I thought.

In summary, not everyone is worried that others are looking....

Friedmud

Re:The Solution Is Crypto

[penguinland]

So why hasn't [PGP] been offered as an automatic part of [email]?

Oddly enough, I'd say that a significant part of it is the chicken-and-egg problem: it's only really useful for cryptography if a lot of people have PGP (note that signing your emails using PGP shows that they're really from you, but does not actually encrypt them; for that, you need to encrypt using the public key of the recipient, and this would require most recipients to have public keys in the first place). For Joe User who hasn't heard of an IP address let alone public key encryption, you'd need some way to automatically set up PGP for him, since he certainly can't do it. and there's no economic motivation for companies to create automatic PGP stuff, since it's not really useful until more people adopt it (as I said earlier), though this is precisely why more people don't adopt it.

On a related note, if you have a PGP key and then buy a new computer, you have to either know what you're doing in order to get your private key onto the new computer, which Joe User also can't do (And if there is a way to automate this process, anyone could write a virus that would use the automated version to steal your private key), or remove your original key and create a new one, which would confuse Joe's friends when their PGP systems suddenly don't trust Joe's email any more.

Sadly, the only way that PGP will become popular is to educate the general populace so that they know as much about computers as we, the computer nerds, do. and although I don't want to admit it, this is never going to happen.

Re:Get on Freenet ?

[blue_adept]

freenet exemplifies what a peer-to-peer internet would be like: a disaster. It's slow, it's cumbersome, and more to the point, it fails to solve a problem a doesn't really exist in the first place. Nobody cares about anonymity at the EXPENSE of speed and convenience, except child pornographers, law breakers, and the paranoid. That's why networks like freenet and ZeroKnowledge ultimately fail.

That's not to say freenet not an interesting experiment. That's not to say anonymity isn't desireable. but please, anyone that's tried it knows it's not a panacea. If you're really paranoid, use a proxy like anonycat, or any of the zillion others. They are more than adequate.

Hmm, well...

[slavemowgli]

A backbone-less Internet... is it just me, or is that exactly the way the Internet was originally envisioned and built? The reasons we moved away from that are purely economical, and until there'll be an economical incentive to move to a backbone-less distributed system again (and, for that matter, an economical incentive to actually make it work at least as well in terms of speed and reliability as the system we currently have), things will stay the way they are now.

The fact that the centralised system of today lends itself to easy censoring etc. is unfortunate, but if you really want it to change, you have to understand why it came to be.

Re:How did this make it to the front page?

[Z0mb1eman]

Oh, chill out.

Not everyone is a networking guru (I know I'm not). I'm sure many people without much networking background have wondered the same thing as the article poster at some point or another, quite likely while reading all the "government/telcos/corporations/Godzilla are going to eat our Internet" stories here on Slashdot. The comments in this story are the perfect place to give these people a better understanding of how the internet works.

This isn't a question that's easy to Google if you don't already know what to look for (in which case you don't need to), and the poster shouldn't have to take a networking course just to get an answer. I would say it's a perfect question for Ask Slashdot - if you don't like the user's ignorance, you could take the time to educate him and the many other Slashdot readers like him with a more informative post.

Re:How did this make it to the front page?

[djSpinMonkey]

Not only does the poster obviously know very little about networking or computers in general (come on... "hosting someone else's packets on my hard drive ... eh.. that's retarded. hard drives are SLOW), but this idea is patently stupid.... I most ignore the trolls about slashdot going to hell ... but this technologically infeasiable and outright rediculous idea should never have made it past the editors. Come'on guys, what is the deal?

Mod parent up +1 Funny. For one thing, the suggestion this guy's ridiculing describes the current architecture of the internet. For another, he's saying you couldn't route packets through your hard drive... because it would be too slow.

Comedy gold, I tell you.

Re:one way it could be done is

[MrPerfekt]

My head just exploded after reading that.

Solving a problem by creating another?

[saifatlast]

Unless you're going to hand deliver your data to the recipient, you will always have to trust someone with it. In a P2P system, the size of the entity with access to your data is smaller, but the number of entities with access to your data is bigger. I contend that it is easier to control and regulate a small number of large entities than it is to regulate a huge number of small entities.

To me, it would be a better use of resources to put regulations into place (and enforce them!)

Pure Wireless Mesh

[Agar]

Seems to me that the biggest risk to individual freedoms is transport over centrally/corporate owned lines.

Why not leverage nearly ubiquitous wireless access points (and possibly ad hoc wireless card settings) to create a completely wireless mesh that doesn't even connect to the Internet at all? This would parallel the development of the original 'net, where it starts as a bunch of island networks that get interconnected over time.

Think about it-no phone lines, no centrality, no existing infrastructure. Nothing to "tap", very hard to track. Even better, no infrastructure so it could be built from scratch. IPv6, anonymizing, encrypted.

Imagine a set of open source tools that take the best features of mesh networks and peer-to-peer, running exclusively over home wireless technology. One package could include a complete set of apps to get "on the mesh" including the routing intelligence, a "secure sandbox" for shared files/web pages, a browser, and caching. Run the package, and maybe at first you only connect to another geeky neighbor-but you don't know which. Check out his home-brew page in the browser, poke around the files he put up. As more people come on line, what you can access increases, sometimes dramatically as networks are interconnected.

(Maybe initially the system could tunnel through the internet to connect disparate networks and gain critical mass. At some point this will always be necessary to get across oceans or challenging geographies.)

Chicken and egg problem? You bet. Realistically, the three p's would drive it, as they do many new technologies: porn, piracy and privacy. But the opportunity is there for so much more.

Speed would suck, sure, due to routing inefficiencies. But consider that the average bandwidth would be at 802.11 speeds: minimum 10Mbps, more likely 54Mbps. If I get 3Mbps on my cable line I'm thrilled. Latency might be high, but no one would be running Quake 3 on this. And wireless tech is only getting faster, while mesh routing and caching technologies are only getting smarter.

I really think that if a truly independent, hacker-run next-gen internet will ever exist, it's going to be over home wireless. The entrenched media companies are too aware of the money making opportunities to let the "free ride" on their infrastructure continue forever (even though it's not a free ride, but don't tell them that). Unregulated spectrum is about the only Free space left.

Use it to create a network that's truly decentralized, owned by the people, and anonymous from the ground up and you can change the world.

Oh, how I pitty them

[MarkusQ]

Imagine, for instance, if Senator McCarthy had been able to steam open every letter in the United States. In the age of ubiquitous e-mail and filtering software, budding McCarthys are abel and willing to do so.

As an administrator of a few reasonably small domains, my first thought was oh, the fools!

You don't want to read every piece of e-mail that comes into even one site, let alone the whole internet. You don't even want to try to write programs to do it.

/dev/null, I tell you, /dev/null! The only sane thing to do with 99% of the e-mail is route it to /dev/null in the most efficient way possible. All else is madness!

You would be better off trying to understand the inner thoughts of a lava lamp then trying to figure out why anyone thinks anyone would buy "farmasuiticals (the 1 U've been lOOking 4!)", let alone ingest them! Or invest in "s+0cks" that are about to "+ake 0ff" based on the say so of a stranger named "Brandice Hornyslut." Or the pointlessly malformed sludge, the server errors from misconfigured machines...if anyone really wanted to hide something they'd be about as well off e-mailing it as flushing it down the toilet--and trying to find it would be about as pleasant.

--MarkusQ

Good God man, you've discovered USENET!

[John Jorsett]

You've described the original implementation of USENET. Participating machines would dial each other up and exchange current traffic. A message injected at one machine would eventually end up in the rec.practicaljokes.hotfoot newsgroup on every participating machine within a day or two, just by this simple machine-to-to-machine exchange.

Absolutely. Encryption, or self-deception.

[Slartibartfast]

The bottleneck is infrastructure: there's no way around the fact that your cable modem/phone line/T1/DSL/whatever winds up at some aggregating point. Wireless is, in a real sense, even worse -- sure, it could avoid said aggregation, but it's wide open. The only true way (and, by the way, the idea behind the genesis of S/WAN) is for encryption to become de-facto. If and when that happens, THEN, and ONLY THEN, will there be the ability to avoid scanning of your stuff by .

Of course, I sure the hell wouldn't put it past the gov't to outlaw encryption. It's not like they haven't done it before.

Uh...IPv6

[NeepyNoo]

'nuff said.

been there, done that.

[Quixote]

It was called UUCP. :-)

Useless and pointless...

[Vexler]

Is /. really running out of news to cover that we have to resort to this kind of "I am not a specialist nor do I really care to do some basic background reading, but here goes" talking points? I see this kind of pseudo-deep-intellectual topics a lot on sci.crypt, where someone would claim to have found a brand-new algorithm, only to have one or several of the following happen:

1) The algorithm gets shot down in about fifteen minutes by several people who really know their stuff,
2) Someone posts, "Oh, this is exactly the same thing as that zippity-zing-zang algorithm that Chuck Dumbo 'invented' some years back. It's completely bogus."
3) Someone posts a follow-up question, and based on the reply given by the OP you suddenly realize that he has no clue whatsoever about crypto design.

It really is not that hard to research some basic, layer-1 information about networking and deduce some fundamental operating principles (as someone already pointed out, one of which is physical cabling). Cisco has plenty of introductory material that even my wife the musician can understand. Do your homework first, and then come back.

Centralization almost unavoidable

[xenocide2]

Look at GNUtella. Years ago, a problem was noticed: some peers are far more capable than others. Search traffic became heavy enough that it was saturating dialup users. This wouldn't have been so bad if the protocol didn't also ask for pseudo anonymity; this led to the networks occasionally dividing in two as a set of dialup users flooded off the net. The solution is to organize the network so that high capacity peers are on the inside, and dialup or otherwise impaired users become "leaves" of sorts. Gnutella2 uses this approach, and this has been added back to Gnutella in some fashions.

The end result of this unequal distribution of resources is that centralization is the most efficient use of them. For the vast majority of Internet users, efficiency and performance are paramount. I hear far more complaint that Bittorrent is slow than that it's centralized or not anonymous. Even if you're willing to discount performance, the price of implementing a peering based system is greater, since it costs to maintain each link. People have tried using wifi to create mesh networks that operate sans "backbone" but this doesn't scale well either. Nor is it anonymous or difficult to tap.

ignorance is so painful

[b17bmbr]

mccarthy, while his methods were excessive, was after communists in the state dept and army. and you know what, there were plenty. we have the venona project as proof that we were infiltrated at the highest levels. and before you defend political freedom, these were people working for the enemy. you konw, the one with 10,000 nukes pointed at us, the same Stalin that had millions of Ukrainians starved to death, that killed many millions more in his purges, sent millions to the gulags, oh wait, duranty was right. those trials were legit.

what makes it even more funny is that bobby kennedy served as mccarthy's right hand man. jack kennedy was a good friend of joe mccarthy, and the real "terror" came from HUAC. but see, that was a bipartisan affair, and well, history is easier just demonizing the republican mccarthy.

as for the NSA thing, monitoring incoming calls is hardly widespread domestic spying. since i'm not a lawyer, i honestly don't know all the FISA details. but amateurish speculation is nothing more than sophistry. oh, as for the history, well, I'm a history teacher.

one last question, would there be as much anti-mccarthyism if he went after fascists? 'cause when you get right down to it, both the communists and nazis were equally evil, equally bent on world control, domination, and destruction. but since uncle joe or chairman mao didn't target those according to their race, i guess it's not really genocide then, eh? the millions dead? just "collateral damage" to be sure.

Re:ignorance is so painful

[Aris Katsaris]

both the communists and nazis were equally evil, equally bent on world control, domination, and destruction. Equally evil? And yet after 70 years of communism in Soviet Union, the Ukrainian/Kazakh/Turkmen/Estonian/etc nations still exist and were not thoroughly exterminated by the dominant Russian nation -- do you think that the Poles would have survived as long under Nazi German occupation? The Jews definitely wouldn't have. The communists were equally bent on "world control, domination, and destruction" as the Nazis? If you mean that they wanted the whole world to follow their ideology, ofcourse -- and capitalists also wanted the whole world to be capitalistic. And democrats likewise want democracy to spread. BUT THAT'S NOT WHAT THE *NAZIS* WANTED. The Nazis didn't *want* black Nazis and Jewish Nazis. The fate of blacks and Jews under Communism in theory would have been the global equality of all nations and ethnic groups (even if practice didn't tend to follow the egalitarian dreams) -- the fate of blacks, Jews and any other "inferior people" under Nazism would be either extermination or *eternal* servitude. In both theory *and* practice. As such, your claims that communists were "equally evil" to the Nazis is a mere inanity -- though ofcourse a very commonplace one in right-wing circles. And yet the genocidal ethnic cleansing began in Yugoslavia only with the uprise of nationalism *after* the fall of the communist regime there. Equally bent on destruction as the Nazis, eh, all Communists? Not so.

why I won't lose sleep over this...

[david_bonn]

The companies that are talking about tiered internet service are mainly ran by pointy-haired people who barely understand this whole internet thing and want to wish it away. Most people, in particular in the most profitable markets, have choices of internet service providers. The ISP who makes a policy change that makes Yahoo!, Google, or Ebay slow will lose customers. Same problem if a particular backbone provider does that to an ISP. The first business to try this is going to learn how easy it is to lose a lot of customers very quickly. There won't be a second time.

I'm even less worried about any persistent efforts by the United States government to snoop on me. Oh, they'll try. But it is doubtful they will ever be very effective at it. I'll admit it is technically possible to monitor all traffic on tcp port 25 that is going through any of the (relatively few) access points that route traffic internationally. With furious effort, you could even store a lot of it -- and think about how much of it would be p0rn spam. Of course, in the modern era, a lot of SMTP traffic is encrypted with SSL, some of it is over VPNs, and some of it might be accessed via other protocols. Some of that email might be accessed through webmail and it won't be immediately obvious how to fish the emails out. Yeah, Yahoo! and MSN might roll over and hand the emails over to a big bad government. But you'd have to be looking in a lot of places all of the time to build an effective police state on top of the Internet we have today. Given infinite resources and incredible competence it might be possible, just barely.

Oh, but did I mention instant messaging (with how many incompatible protocols)? Did I mention online fora?

Resources and competence seem to be rare goods in the U.S. Government these days. Why should halfhearted snooping be somehow special?

Remember, this is the same government that didn't connect the dots on 9-11.
Remember, this is the same government that connected dots that weren't there in Iraq.
Remember, this is the same government that botches monster iT projects (the FAA and the FBI) all the time.
Remember, this is the same government that still hasn't translated all of the documents captured in Afghanistan.
Remember, this is the same government that did a heck of a job on New Orleans.
Remember, this is the same government that hasn't captured Osama, and took years to capture someone hiding in North Carolina.

Why Not.

[darqchild]

-The complexity of the routing tables. Although people complain that we are running out of IP address space, this isn't exactly true. The problem is in badly fragmented IP address space. That is to say that the route tables of our core routers that join the backbone providers have grown to be huge. There are a whole pile of class C networks (254 hosts each) that the IANA is trying to claw back so they can be consolidated into larger /16 and /8 CIDR networks.

-BGP AS space. Due to what i can only assume was poor foresight, the AS# used to identify BGP "Autonomous Systems" (Corporations, and entities that use BGP to exchange routing information with the backbone providers) is a 16 bit value. So there are only ~65K numbers that can actually be given out.

-Complexity of configuring these routing protocols. It's rocket science, plain and simple. A misconfigured BGP router will not work, and may even disrupt traffic over the rest of the internet. If anyone was allowed to broadcast any BGP route without the consent of all their peers and a pile of red tape, i could advertise a route to 24.0.0.0 and half the internet would disappear for a good number of cable-broadband users.

-Required bandwidth, and latency problems. The current top-level backbone providers have many millions of dollars worth of equipment and high-speed point to point connections to keep the number of hops for each packet to a minimum. They have the capacity to push more traffic than you'll use in a week down their wan links every second. This is a vast improvement over a pile of 56, 1024 and 3068 kilobit connections that would be meshed together in a distributed model.

Re:Why Not.

[Tim the Gecko]

I was pretty sure that there were more than 65,000 BGP AS numbers in use

No - here are the details of the 40,000 or so AS numbers handed out by IANA. There is also a set of weekly statistics posted on NANOG which shows that 21,484 of these AS numbers can be seen in the global routing table. Only 8,867 of these guys advertise a single prefix, so to get 181,747 routes there are a lot of ASes advertising multiple prefixes.

You Are Not A Network Professional

[roderickm]

First, the stated privacy concerns are no justification for changing the underlying infrastructure. If you're genuinely concerned about privacy, then start encrypting everything you put on the wire. Use anonymizing services.

Secondly, network geeks in general do not grok the economics of the internet on a national or global scale. Without statistical multiplexing and large economies of scale created by the "backbone providers" vilified in the original post, your internet access fees would not be as affordable as they are today. Without large service providers, your connectivity would not be as robust and reliable as it is today.

Finally, large-network interconnection is as much an art of negotiation as it is a science of traffic exchange. Each commercial network relies on access fees to remain solvent, but universal access to the internet requires at least a few large players to exchange traffic. It works best network-wise if this exchange is settlement-free and frictionless: routing protocols get to do the jobs they were designed to do, and bits fly directly to their destination networks. However, networks often want to be paid for such peering, on the basis of unequal exchange, network size, stability, POP count, etc. Adding this "friction" to the creation of network peers balkanizes the net somewhat, and arguably increases stability, but it prevents a rich, dense routing mesh that would be ideal for network efficiency.

Just imagine how wonderfully the internet would work if every AS peered with every other AS in a 50 mile radius. Sure, smaller players would still need to buy transit bandwidth, but two businesses in the same town wouldn't need to send traffic to a coast just to communicate. The optimal way to reduce the need for "huge backbone pipes" (a brutal oversimplification, btw) is more dense interconnection and more direct routing that would result. The drag on such progress is economic and political, not technical.

Lord of the ... Tiers?

[bi_boy]

Control them, and you control the net.

One Tier to Rule Them All. One Tier to Find Them. One Tier to Bring Them All and In The Darkness Bind Them.

Yeah I know, redundant, I couldn't resist though.

Re:Lord of the ... Tiers?

[Onan]

Oh, it was necessary.

I just can't believe you passed up the opportunity to end with "..and in the darkness, BIND them."

Re:Maybe Possible and Makes Sense

[jibjibjib]

no, it doesn't. What you are describing is a centralised tree network, not a small-world network. A network such as the one described in the previous post would not have a 'central office' from which connections are distributed. It would instead have mostly local connections between neighbours, which is *completely different* to the current internet or phone system.

most IGNORANT Slashdot story ever

[puzzled]

Wow, its as if the drooling wireless fanboys suddenly discovered life beyond an IP address assigned via DHCP. Please pay attention, children ...

    The internet is composed of 'autonomous systems' - each autonomous system or 'AS' has one or more netblocks of a /24 or larger in size. Each AS connects to at least one other AS, makes at least one netblock available via BGP, and thusly the internet is stitched together. Find this shocking an incomprehensible? Try this

telnet route-views.oregon-ix.net

  follow your nose through the login procedure, then type 'show ip bgp [your IP address]' and see what it says. Oh, if your IP address is 192.168.x.x, 10.x.x.x, or 172.16-31.x.x and you put that in please step away from the computer now and ask someone with a clue for help.

    I mean really - *this* is a frontpage story? I swear I'm going to auction my low Slashdot ID number on Ebay one of these days and alias this site to memepool in my hosts file.

OK, take these steps

[puzzled]

  Maybe I'm getting grouchy in my old age - see parent for details. This is how real men connect to the internet:

  There are three ISPs in the world - Sprint, UUNet, and [other]. Get on the phone and order a T1 from one of the two real ones. They'll get your payment information and then someone will ask how many IP addresses you need. Tell 'em you want a /24 (256 addresses). They'll ask why, you tell 'em you're going to multihome.

    Go to ARIN.net's site. Figure out how to get yourself an autonomous system number. Call up the other ISP you didn't originally order from and get a circuit from them. No IP addresses required, we'll just use the block from ISP 1.

  Assuming you're using a Cisco box do the following:

    router bgp [your AS number]
        network [your shiny new /24]
        ! UUNet
        neighbor yadda yadda AS 701
        ! Sprint
        neighbor yadda yadda AS 1239

    And *poof*! Your little /24 is now globally visible via two different ISPs. Yank the T1 to one of then, life is funny for a bit, then you're running like nothing ever happened.

    Take this little story and abstract it a bit - there is no 'backbone' to be found on the internet, just a web of large carriers with all sorts of peering agreements with each other. This won't happen at the home DSL router monkey level, but the diverse internet the asker speculated about already exists and happens to be pretty resistant to fools trying to monitor it.

Anybody remember FidoNet?

[birge]

Grassroots nationwide network made up of people connecting to nearby people via modem. Granted it took a day or two for mail to make it across the country, but it was pretty cool given that it was truly decentralized and done entirely by hobbyists.

Anyway, I think it's a moot point. Who cares about the topology of the internet when you can just encrypt everything? Backbones are great. Best thing is to use the fastest and most robust network topology, and let security be handled at the application level.

For the moment...

[abb3w]

Unless some higly specialised entities have developed quantum computers and kept it a secret, they won't be able to break it in any time frame suitable for mass communication snooping.

True; but if various corporate proposals go through, your encrypted traffic might travel cross country at sub 56kbps rates with multi-second latency. Which does bad things to a torrent.

Mind you, this still won't stop file sharing. As an example of the alternatives: someone in my apartment complex has a non-internet wireless access point, "Blacknet". It's an "open" network, DHCPing on a 10/8 space. Any DNS query resolves to the IP address of a single server, "OneTrue.blacknet."; and yes, that's the whole FQDN; any traffic to any other IP and any DNS name routes to and is intercepted by OneTrue. OneTrue's apache server redirects any URL not using OneTrue by name to OneTrue's home page. OneTrue also speaks IMAP and POP (any account name and password accepted, any mailbox you check has only one email message directing you to http://onetrue.blacknet/ telnet and ssh (assuming you're stupid enough to accept the key....), and even gopher. On the web server proper, there's about 200GB of MP3's, about 3TB of movies (uncompressed DVD ISO). They have a submissions page if you want to upload MP3s. An "about" page claims the server has over 10TB of space. Games? There's... er, NetHack. For all of the OSes listed at Nethack.org; hm. "We'll put up more games once we get back with the Amulet of Yendor."

They're fucking nuts. Not that I have room to complain, mind you....

Stop government regulation (and try JXTA)

[nervesystem]

As other posters have said the Internet is already (in the industrialized countries) a well connected mesh of peer networks. It's true that traffic flows through the tier one provides but that's only because they provide the best route to where ever your data is trying to get to. If a network provider stops routing traffic or starts censoring or port blocking certain applications then it's your job as Joe consumer to pressure your ISP to not use that providers backbone.

The real threat to the Internet as we know it is government regulations designed to "level the playing field" between VoIP and IPTV vendors and old line PoTS and Cable monopolies. The old time monopolies got their status from the Government by agreeing to a whole raft of "universal service" and other government mandates. These mandates sound great but really just drive up costs and slow innovation. The monopoly companies want to hoist these old rules on Internet providers knowing it will kill their businesses. A good example is trying mandate E911 and WireTap features for VoIP phone companies. Cable companies are getting in to the act to and saying that phone companies shouldn't be able to compete with them by offering IPTV because Telcos don't have the "universal access" rule of having to provide TV to everyone in a franchise area. The monopolies also claim if you get too many providers trying to offer service in an area the streets will be torn up all the time which is also a bogus excuse. Everyone should have access to public rights of way and the cities should just set rules about when and how long streets can be disrupted to cause the least annoyance for people. It's the phone and cable TV monoplies who today wine and dine the cities to let them tear up the streets anywhere and any time they want.

The RIGHT (tm) solution is to drop government regulations and government sponsored monopolies and leave it to the free market to innovate solutions. What right in a free society does the Government have in be involved with any communications business (except as a paying customer)? If cable companies can't compete with IPTV by offering CableTV at a decent rate then let them go bankrupt and a let a company who can do the job buy up their network and make it work. Same goes for phone companies, if no one wants to buy over priced phone and T1 lines from them then get out of the business and let someone else manage all those pretty copper strands. I'm sure there are plenty of smart companies who can use them for phone, Internet, TV, and who know what else.

On a related note, there is one major choke point in the Internet and that's the stupid DNS system. Just FYI, the internet (IP, UDP, TCP, BGP, etc.) will work fine with out it. All it does is take a server name everyone can remember and gives you back the right numeric IP address (66.35.250.150) for that server (ok it does a few more things but that's the basics). Anyone is free to invent a new efficient decentralized network address to network number system to replace DNS. An example of a very cool system that does just that is called JXTA (http://jxta.org) from the good people at Sun Microsystems. It's billed as a P2P protocol and collaboration system but it is also a beautiful re-imagining of the Internet sans DNS.

What People Seem to be Forgetting...

[eno2001]

...is that the poster didn't just talk about privacy, but also about media control. While encryption might handle the privacy angle it does jack squat for getting an unpopular message out to everyone over channels controlled by people who think the message is detrimental to them. Especially if your web host or ISP is told that your message is "illegal" in the next few years. I live in America where it's getting harder and harder to get the truth out to people via mainstream channels. And now we've got politicians trying to shut down bloggers because the bloggers disagree with their views. Political dissent with the right wing in the U.S. is slated to be a crime before the next election. THAT'S the more important issue here. The only way to fight that battle is, sadly, with lots of money which the sane people in the U.S. don't currently have a lot of. To the remaining REAL Americans, media control is a HUGE issue. The wrong people are controlling the media today and the Internet is largeely becoming just another form of media for them to rule over.

(OT) Re:ignorance is so painful

[cgenman]

Japan was attempting to destroy the US as well. No doubt there were Japanese infultrators amongst the citizens of the US... some must have have been in positions of power in their communities.

But that doesn't justify taking the lives and families of Japanese Citizens of the US and throwing them in concentration camps. That does not justify locking my grandparents up like criminals for years, kept away from their kids.

McCarthy didn't just go after traitors. He went after communists, people with alternative sexualities, liberals, those that believed in social support, those that felt capitalism needed work, and anyone that anyone was willing to name to get themselves out of trouble. Just like being ethnically japanese made people potential traitors in WW2, being of the opinion that pure capitalism is broken was enough to get you thrown in jail. Even agreeing with Adam Smith that the pure capitalist system eventually breaks down was enough to get people blacklisted, thrown out of work and schools, careers and futures taken away from them. And remember, Social Security was considered a liberal, communist thought. There is a lot of ugly, pointless history there.

And its happening again. Now we're throwing people in Guantanamo if we suspect them of being a terrorist. And a terrorist is anyone who disagrees with the war on terrorism. Being a darkie, of course, doesn't hurt, just like racism played into our concentration camps in WW2 and our ideological purge by McCarthy.

You're a history teacher. You should know better. If you can't see the connection, history is most assuredly doomed to repeat itself. And who knows who it will be next time: lots of countries have purged their intellectuals.

we had that kind of network...

[sednet]

Is it possible to create an internet that relies instead on peer-to-peer connectivity? How would the hardware work? How would the information be passed? What would be the incentive for average people to buy into it if it meant they'd have to host someone else's packets on their hard drive?

we had this type of backboneless internet, once upon a time, operating under various names:

• UUCP
• FidoNet

With one or two minor changes, yes.

[jd]

The question is basically a re-statement of the original ARPAnet design, you are correct. However, to be absolutely true to the question, you'd need two additional stipulations.

First, to be effective, all network connections would need to be fairly fat. A tiered Internet is designed along the same sort of design philosophy as a "fat tree" - low bandwidth at the work-node level, massive bandwidth in the middle. A tierless Internet, particularly one that supported enough multiple paths to be useful for robustness and decentralization of control, would need ALL connections to be much fatter than they currently are. You'd need gigabit to ten gigabit pipes between the majority of machines to be useful.

Second, you can't use the design strategy of bordered autonomous clouds, linked by a backbone, because you'd have no backbone. With no borders, you can't use internal and external routing protocols, as there would be no "internal" or "external". Besides which, they mostly suck when it comes to massively meshed networks where individual connections are unreliable and potentially mobile. BGP, OSPF - you'd need to RIP (yeah, bad pun) them out and replace them with an ad-hoc mesh routing protocol that supported mobile IP and NEMO. The complexity would be much higher, particularly as software packet switching and software routing are CPU and bus killers, which means an optimal path would need to figure in the density of traffic in a fairly sizable part of the mesh. Modern architectures just aren't built to handle such a design, but that would not stop you from building an architecture that COULD support it.

So, (1) yes it is possible, but (2) not effectively with the existing infrastructure or existing PC designs, though (3) both of those problems are solvable.

Wait!

[Ethan Allison]

If we had no backbones, how would we get across the ocean?

Even meshes are no solution to state interference.

[Ivan Matveitch]

Since the state would simply ban any potential technology of circumvention---ie, the police would arrest you for illicit wireless networking. Though, on the bright side, there are probably some authoritarian states that are now institutionally incapable of carrying out such a policy.

Non-corporate Tier1, that's how.

[HBK-4G]

P2P isn't the solution. The solution is to take a significant majority of the backbone out of the hands of corporations that control them. Corporations will bend to government influence, just as governments bend to corporate influence.

By creating a non-profit organization whose sole purpose is to enhance and extend the internet backbone, you've solved the problem of petty ownership and government blustering. Funding would be an adventure, but it's been done by lesser qualified organizations. And no more Level3-Cogent spats!