http://www.cbr1100xx.org/temp/jul/nofishing.jpg

I don't know if it's possible to make any change right now that would be logical and proper. There is so much political and financial pressure and the stakes are high. I really don't think there's a problem now, and don't foresee one in the near future with things the way they currently are. I'm open to seeing examples of why things need to change though.

Ok, let's get closer to the original post. The question was about the structure of the internet. I think we're all agreed that the physical structure is very robust and incredibly redundant. Upgrades to the physical structure take place all the time with no discernable interruptions to users. So what needs fixing? The logical structure dates back to the very begining of the internet. Back then, 4 billion addresses seemed incomprehensibly large. Today, more than that number of addresses could be used by China alone.

The problems now are:

1. Far too few ip addresses. This requires dynamically assigning addresses, which because they're more difficult to track, make it far too easy to create problems 2 and 3.

2. Spam accounts for most (almost all!) of the email sent. This isn't just a problem because it's annoying, it also uses an enormous amount of bandwidth. The added expense gets passed along to consumers, who also have to deal with slower internet because of it.

3. Viruses and spyware affect all of us. As Mac users, we may feel that we're unaffected, but we're not. We may never have to deal with a virus or spyware on our systems, but most of the institutions we deal with are constantly wasting money dealing with both, and they pass the costs on to us.

The current system costs more and delivers less because of these problems. When the switch to IPv6 is completed, it will be possible to significantly reduce both. Whether or not they are actually reduced will depend on who is in charge. That's why discussions about the future of the internet must always center on the political aspect. None of what we'd all like to see is technically impossible or even difficult. Getting a consensus is.

There really aren't even 4 billion address spaces in IPv4. Remember that the founding fathers, never anticipating the exponential explosion of internet use, made some decisions that we are paying for now. The 32-bit address space is bad enough, but their original ordering in octets has caused problems too.

Originally, networks were classified as A, B, C, or D, where an A spans the first octet (a /8 network), a B, the second octet (a /16), and a C is a /24 supporting only 255 addresses. D was a special 1.1.1.xxx used for multicasting. To make matters worse, IP address space was allocated on the basis of requests for it rather than on need. If I recall correctly (I used to teach there), MIT, for example, has all of 18.xxx.xxx.xxx - they asked for and got an entire A address space as one of the founding institutions. The octet grouping is also problematic. A /24 (just the last octet) is restricted to 254 hosts - too small - and a /16 (the last two octets) is 65,534 hosts - really too large in most cases.

To compound matters, quite a few of the available numbers are Private Addresses. These are addresses that a router will not pass on to the WAN.

10.0.0.0 - 10.255.255.255 is a private network range often used in corporate networks.
172.16.0.0 - 172.31.255.255 another large private set.
192.168.0.0 - 192.168.255.255 - the common set used in LANs and small routers.
0.0.0.0 - default route
127.x.x.x - loopback address - a whole A address space. Zeros at the end refer to the network
x.x.255.255 is a broadcast address to all machines on x.x.

I haven't studied IPv6, but in addition to more space, it presumably allocates that space for a continuing explosion. The most important factor in this expansion will definitely be a realistic allocation scheme or the IPv4 problems are just delayed for a while.

It is tempting to go after all the flame-bait that there is in the thread now, please be brave and resist.

Right, Voldenuit. :) There are lots of red herrings and straw men popping up as well, which complicates discussion.

--------

One question I've had is whether the net will be able to handle the traffic that's sure to increase through the years, especially if people begin to stream more and more movies. So far, things seem to be working out (re. bandwidth), but that could change. I'm sure the various providers of backbone services are aware of an are planning to accomodate somehow.

Anyone have any thoughts or concerns about the expansion process?

cwtnospam: That's why discussions about the future of the internet must always center on the political aspect. None of what we'd all like to see is technically impossible or even difficult. Getting a consensus is.

OK, but it's happened so far with a rather "laissez-faire" attitude re. control from politicians. Why could it not continue to happen that way? Phasing in new technology and protocols has gone on all along. What's so different now?

Well there is of course Internet 2, but that is primarily restricted to the research institutions that lost the original Internet, and don't plan to make the same mistake twice. ;)

What's different now is that the internet is much more commercialized than even a few years ago. It used to be about information, and now it's more about selling things. As the internet becomes more and more capable, it becomes far more likely that individual users will be overwhelmed by marketing messages. Can you imagine what would happen to your inbox if all the current spam blocking techniques being used by ISPs were to be rendered even 10% less effective? You wouldn't have enough time in your day to sort through it all.

Communities in the real world have passed laws to limit billboards and glaring neon signs because the signs threatened the quality of life in those communities. Now it's time to do the same with the internet, before its usefulness is negated by its pitfalls.

Wow. Talk about a political question! Or are you asking about laying fiber and installing servers and routers?

Good to see that we finally come to have a common appreciation on the conduct of a (slightly) political thread ;) .

Well at least in Europe, I am not worried. During the dot.com bubble years dark fibre has been buried at a scaring pace and combined with advances in modulation, I can't see a shortage coming anytime soon now. It's just a matter of putting another card in some massive Cisco or Juniper router.

In France, one operator, free.fr, has even started combined VoiP, TV and internet service over ADSL2 lines at speeds up to 20 mbit/s for 30 euro/month.

NightFlight (vol_de_nuit) is right, there's tons of fiber available - too much for current usage which was among the sparks that ignited the dotCom crash. But, and I don't mean to make this political again, much of this capacity is owned by the countries it's in.

ASIDE:

I did a consulting job in Chile some years ago at the behest of Cisco and was surprised to discover that the government of Chile had "wired" the country so that the internet reached every school in the country. They are internally well-connected. Their problem back then was limited (and slightly unreliable) bandwidth to the USA from there because they didn't control the pipes in between which then passed through Peru, Columbia, and finally reached Texas. Almost all the equipment was either NEC or Cisco, by the way.

Looking at a traceroute there now (3.5 secs) it goes from Nova Scotia to Australia for DNS, back to a switch in Toronto, then to the megacomplex in Fairfax, VA, on to an Australian switch, and finally to Chile. Roundabout, but entirely dependable.

Thanks for the constructive posts et bien le bonjour aux francophones ;) .

I think novascotian brings an important issue into the discussion:
Although potentially touchy, the "digital divide" certainly is a problem and whatever contributes to a wider, unfiltered access to the wealth of information available through the internet to more people certainly is a Good Thing.

From Europe, I have about 200ms latency to apple.com and only 70ms more for a target in Chile which is not that bad after all.

And to support the "laissez-faire"-attitude in net.governance, let me explain why I don't think that any of the three points raised by cwtnospam are to be addressed by IANA/ICANN:

1
For now, there is still some slack and there are quite some unused class A networks to be put into RIR allocation pools. And I am extremely worried by the legal initiatives in Europe trying to get ISPs to store IP-customer-logs for an entire year. Why not require the post-office to keep track of a year worth of your mail ?
IPv6 has a hen-and-egg-problem for sure where everybody waits for someone else to move first, but at least it is there, ready to be used once the real IP-shortage gets started.
BTW, China will start to use IPv6 soon, check out the article I quoted in post #23 in this thread.

2
Spam clearly is a problem. However, the conviction of Jeremy "spam-king" Jaynes to a nine years prison sentence under Virginia state law shows that the legal instruments exist to fight it - sensible territorial laws are all it takes.

Also note that the major proportion of spam these day is originated by 0wned Windows machines being made part of a botnet. Looking at the responsibilty of a corporation producing an OS that will, when connected in its out-of-the-box-state to the internet has only minutes to live before it gets hacked seems to be a lot more sensible to me.

As it is unlikely that the small proportion of morons who actually fall for spam mail will shrink to the point where it does no longer make sense to spam, technical solutions will be perfected to deal with it or other concepts such as

http://cr.yp.to/im2000.html

will replace the naive SMTP of today. (check out the guy behindd the idea, djb is quite a personality).

3
Again, while the basic statement that malware sux rox is true, efficient action should probably look at the vulnerable OS rather than pro-active port-blocking or other misguided band-aids. Corporations may be more or less ruthless, but they certainly try to not lose money on computer systems with a high TCO. Which is good news for Apple.

2 and 3 are annoying criminal activities and can be dealt with using the legal system of the country where the damage occurs. They are part of the not-exactly-perfect life as much as shoplifting. Yes, you pay higher prices because there is "shrinkage".
So, death-penalty for shoplifters ?
Hopefully not...

But increasingly less so for individuals; my ISP catches some, Spamfire catches most of what's left (but hasn't yet figured out that I don't understand any language not written in the Roman alphabet with a few accents thrown in), and Eudora sorts that out for me, so I spend only a minute or so per day dealing with it. Similarly, pop-up blocking is effective for web sites that don't serve their own ads.

The problem then is really on the dark fiber lines where gigabits/second of bandwidth are devoted to spam that never actually reaches anyone and goes unseen, and except for the expense to us, it's not our problem. Mass mailings do have their uses - I'm often informed of software upgrades and new products by this path (though often after I've seen them in an RSS feed).

Given, however, that there is currently such an abundance of dark fiber capacity, no one is really paying much attention to the spam problem (with the exception of Virginia, I guess); they're leaving it to the end users to deal with it. When that trunk and giga-switch capacity begins to be strained, however, THEN, the owners of IP backbone and the providers of the switches that control it will jump in to do something about spam. I doubt very much that it requires govenment regulation to get there - all it requires is a cost to the spammers that makes spamming unprofitable.

The latter, and some of the responses are suggesting that the capacity is already there. That's good to know.

True, but that method has already shown itself to be far from efficient, and relying on it invites the kind of government intervention that so many seem to abhor!

FWIW, I never read anyone on this thread objecting to legal action against people who violate laws pertaining to use of the Internet. Countries making laws about spam, pornography and even net commerce isn't the same kind of intervention as "overseeing" or "regulating" the ongoing development of the Internet. The consensus of most participating in the discussion seems to be that this has gone quite well so far, and there's no reason to "fix" what isn't even "broken."

I didn't say they objected to legal action, only that relying on it can and probably will result in just what they do object to, since governments find it difficult to pull their noses out of anything once they stick them in.

I wouldn't say the internet is broken either, but I don't believe that in the long term it can't continue to grow the way it has been because the abuses it allows will outweigh its benefits at some point.

Great response!
 

Wow!
I just got back from hiking and fishing in the mountains; posted the tread just before I left. Really glad to see the interest in the topic. Looking forward to reading the posts in detail and checking out some of the links.

:)

Another interesting illustration how ICANN is not independant as it should be and the US WSIS delegation never stopped affirming, is the case of the .xxx domain and how it got killed by right-wing religious pressure groups.

Let's keep things apart here:
It may or may not be a good idea to have such a domain. There are, as you can see from the document I'll refer to later, also people in favor of the .xxx domain with the afterthought of then forcing all "sexy" content to go exclusively there.

That's not my point.
My point is:
Us-american lobbies and day-to-day political opportunity get to decide what happens to the internet at large, rather than a neutral institution with a net-wide horizon and a long-term vision like it is supposed to be.

Here's a short article:
"On Friday, May 19, 2006, ICM Registry announced that it would file a reconsideration request with ICANN. ICM Registry applied for the .XXX top level domain from ICANN and was turned down May 10 following pressure on ICANN from the US government. It is also filing a judicial appeal under the Freedom of Information Act to challenge redactions and omissions from the internal US government documents released to it under a Freedom of Information Act (FOIA) request.In connection with its judicial appeal, ICM Registry released 88 pages of documents obtained under the FOIA showing how the U.S. handled its application. Even with the major redactions, these documents show how US supervision of ICANN was influenced by domestic political pressure. They leave no room for doubt that the US altered its policy toward ICANN in response to this pressure, and that it actively worked in tandem with ICANN to conceal the nature and significance of US governmental influence over ICANN from the media."
http://www.internetgovernance.org/ne...foiaxxx_051906

And here are e-mails and other documents showing how things happened behind the scenes, obtained, with some redaction, under the Freedom of Information Act by the would-be-xxx-registrar:
http://www.internetgovernance.org/pdf/xxx-foiapage.pdf


Even if you disagree, please, be good to this thread, try to go for articulated dissent. There has been enough shouting already.

An argument for international control
 

.
Those of us who are concerned about the future of the Internet are in good company. There is an interesting article on BBC about Sir Tim Berners-Lee, the inventor of the Internet.

He is concerned, amongst other things, about US efforts to charge for different levels of online access -- i.e. a two-tier Internet. Tim Berners-Lee, director of the World Wide Web Consortium, insists on an open mode.

Interestingly, he points out, Microsoft and Google want legislation guaranteeing the same. (Ref: the Net Neutrality Bill currently before the US House of Representatives.)

So, who disagrees? Who wants a closed model? Well, telecom companies for one. If they succeed in their efforts, giving better access to those willing to pay, they would essentially become gatekeepers for content as well. Why? Because one group of content providers that might be prioritised is Web TV, and they have the bucks to pay, and TV audiences may be equally willing.

Personally I find this sort of discrimination very questionable.

Should the Congressional bill fail, it seems yet another excellent argument for yielding control to an international body, and not on controlled by just one nation. It seems more likely that we can then avoid what Tim Berners-Lee fear would be "a dark period for the Web".
.

MacRumors headline on Network Neutrality
 

.
MacRumors has clearly found Network Neutrality important enough to headline the issue on 26 May. For a front page topic, they have a rather lengthy article about the proposed Congressional bill. (There is more background info on Network Neutrality on Wikipedia, including an in-depth brief of the bill now before Congress.)

It will be very interesting to see if it passes both chambers -- and finds a willing Presidential signature. That would be a strong statement, making it much more difficult to move in the wrong direction.
.

Just to make sure there's no misunderstanding:
Sir Tim Berners-Lee invented the www, not the internet (nor did Al Gore...), which is the "enfant naturel" of the US military and the hippie generation and something about as all-american as it gets.

Network neutrality is an important issue, understanding that everybody being able to connect with everybody else regardless of marketshare is the basic paradigma that made the internet such a wild success.
The very fact to bring up a debate to question that principle is scary.