Slashdot Mirror
National Projects Aim to Reboot the Internet
iron-kurton wrote with a link to an AP story about a national initiative to scrap the internet and start over. You may remember our discussion last month about Stanford's Clean Slate Design project; this article details similar projects across the country, all with the federal government's blessing and all with the end goal of revamping our current networking system. From the article: "No longer constrained by slow connections and computer processors and high costs for storage, researchers say the time has come to rethink the Internet's underlying architecture, a move that could mean replacing networking equipment and rewriting software on computers to better channel future traffic over the existing pipes. Even Vinton Cerf, one of the Internet's founding fathers as co-developer of the key communications techniques, said the exercise was 'generally healthy' because the current technology 'does not satisfy all needs.'"
Just like IPv6 is going to be implemented... someday. It will never happen. I guess someone needs to bring it up that, boy, it sure would be a great idea, but frankly it will never happen. The Internet is so much more than just the US, there's no way you can have it scrapped. As is with most things in this world, it will continue along on this current path, and maybe something will be built along side it (eg Internet2 or whatever that University network is called) and eventually switched over, but you can't just scrap it.
I think the last paragraph is disconnected with reality, but the second paragraph makes a good point or two.
I wonder if I use bold in my signature, people will notice my posts.
Stanford University states their research program can be characterized by two research questions: "With what we know today, if we were to start again with a clean slate, how would we design a global communications infrastructure?" and "How should the Internet look in 15 years?"
A new internet architecture such as proposed will open vast new markets and endless business opportunities - in short - a potential gold mine for the seven industrial sponsors. The fear is that the Stanford research program will trade off attention to social and political issues for expediency in the impetus to get the new infrastructure up and running quickly.
How do we ensure that those questions don't get switched around to, "...if we were to start again with a clean slate, how would we design a better conduit to more efficiently funnel revenues to our sponsors?" and "How should their profit margins look in 15 years?"
See my blog "The Internet is Broken" for an answer.
The biggest shortcoming of the current internet (to me) is that anonymity wasn't designed in from the ground up.
Hopefully, this "next big thing" will be designed so there is no information (like IP address) that can be used to trace an internet persona to an actual person or geographic area.
I certainly agree with this. I have heard of plans similar to this to force people to upgrade, and at the same time accept DRM loaded software. When we look at this, the current network protocols work just fine. TCP actually works pretty well, its not really something that is TCP does have a few limits built in I believe (i believe it is the window sizes), that might make it a bit limited with extremely high bandwidth connections, but those limits are far from being reached and it is more than adequate for video, audio and other things. New versions of TCP can be developed if needed, and those can be made to work side by side with the older protocols. To say that everything has to be scrapped is just beyond ridiculous, since what upgrades which are needed can be made incrementally and with backwards and forwards compatability. It would be completely wasteful of resources, since most router and equipment actually works pretty well. If needed it can be upgraded, but to say everything has to be thrown own, when it is functioning just fine, is beyond ridiculous. The current internet is working well, and with proper management it will continue to do so.
One of the major problems with IP6 is the lack of really much of an expectation that it will need to interoperate with IPV4 for a very long time. One problem is, no one will upgrade to IPv6 since there are few websites that use it, and since no one is upgrading to IPv6, few websites are inclined to provide it. ISPs, with newer OSs if IPv6 is autoconfiguring, the users computer will automatically configure itself for IPv6. But to expect all ISPs to adopt IPv6, especially before IPv4 address space runs out, is just beyond arrogant. There has to be expected that IPv4 ISPs will be online long after IPv4 address space is maxed out, and IPv4 systems will need to be able to access IPv6 systems coming online then. Ipv6 accessing ipv4 hosts is simple, make ipv4 a subset of ipv6. One of the major problems is IPv4 being able to access IPv6 hosts, new hosts can be given v6 and v4 addresses, but this means that the address space problem has not been solved. But ISPs can be expected to continue using only v4 with some existing users, for some time after v4 address space is exhausted. There are ways for v4 to access v6, through a concerted effort of DNS servers and routers. When a v4 peer askes the local DNS server for a the IP address of a server which is v6 only server, the DNS server will return a fake v4 IP address to the v4 peer, and tell the router (which would have connections to the Ipv6 net) to to route all packets going to that fake IP coming from that v4 peer, to the IPv6 destination, converting the packets to Ipv6 as well. If a IPv4 peer wishes to access a Ipv6 peer by Ipv6 address, a neat trick also using DNS would be used, a special ip6 top level domain would be created, and ipv4 clients could request Ipv6 addresses by specifying ipv6 addresses as subdomains as of this ip6 tld. such as: 2222.2222.2222.2222.2222.2222.ip6 A portion of v4 address space needs to be set aside for this scheme for use for the fake IP addresses. Proxy servers could be provided by ISPs to convert Ipv6 hyperlinks to hyperlinks using the ipv6 tld notation. Newer web browsers could automatically do this for the user if they are on an ipv4 only network. Problem solved! This would require no changes on the user end, and the ISP could even use 6-over-4 to connect their routers to ipv6 networks even if they are not directly connected to an upstream ipv6 provider.
I hope this proposal fails. They probably know it will fail but this is a "make work" project -- just a method of researchers to spend grants so that they will get further grants like governments do at end of fiscal.
t hreadID=190&messageID=26576&start=-1
When they talk about "The internet" they mean layers 4 and 5: IP and TCP (or TCP/IP).
[Points about the internet]
- The author of that article knows nothing about the internet; not once did he mention any of the layers.
[[Important]]
Just because the internet is old does not mean that it is by any means bad; in fact, for most users, this "older version" is better.
- IP was formalized back in the 1980s; it was designed back when memory was expensive; it was designed when every bit and byte counted. A new "version" of the internet would in fact be much slower than the current IPV4. IPV6 for example has more over head than IPV4.
Read about the formalized protocol here
http://www.faqs.org/rfcs/rfc791.html
The paper may seem a little complex; however, realize that the purpose of IP is for simply routing messages so that one machine on the internet can reach another
-- Final comments --
The only service which the "old" internet does not offer is functionality for quality of service (technically it does, but those bits are not used).
And this is a good thing. No quality of service prevents ISPs from gouging their users. How would you like it if your ISP implemented IPV6 and then said "Oh, and if you want your latency below 200ms that will be an extra $20/month).
That is the type of functionality they are looking to add into the "new internet". Anything else can just be built ontop of TCP.
---
Also, to clarify the IP running out of addresses issue: TCP/IP supports 4 billion address. Only 1 or 2 are in use. Why is there a shortage problem you ask? Because of the method in which IP addresses are assigned. IP addresses are assigned in classes of A, B, and C (do further research for understanding this).
For example, Stanford university has more IP addresses than all of china.
http://news.com.com/5208-1028_3-0.html?forumID=1&
A simple bandwidth guarentee system is at the top of my wishlist.
How we know is more important than what we know.
Yes, I can. It might not be recent or entirely relevant, but the entire US rail network south of the Mason-Dixon line was converted from broad 5ft gaguge to the "standard" 4'9" gauge that was used in the North on May 31 1886. The work was completed in less than 36 hours.
No matter how you spin it, that's pretty darn impressive.
-- If you try to fail and succeed, which have you done? - Uli's moose
For the most part, I don't think spam is the Internet's fault. I think superfluous messages are the cost of ridiculously cheap and convenient communication. Spam a pain, but not worth locking the Internet down to combat it IMHO.
I guess you've never heard of the whole net neutrality debate. Or for that matter the DMCA (a means for certain companies to selectively banish whatever they want from the 'net, at least temporarily). Or domain name trademark disputes. Or "great firewalls" filtering entire nations' net access, with the aid of US companies. These are real issues. It's ridiculous for you to claim that the powers that be aren't clamoring for more control over the Internet when they've been doing so for at least the last 10 years, with some success. You think now they're suddenly satisfied? Pass the bong.
What is really needed is widespread adoption of encryption; this would prevent the hoards of greedy and evil entities from pushing "solutions" to problems which don't actually exist. The purpose of the network should be to move data, not to enforce policy, or spy on people. Things such as VOIP are recent enough that they should never have even existed in an unencrypted form. At this time, any fundamental redesign of the Internet will likely only make the situation worse.
Thankfully, this is a problem that can be solved at the edges of the network. If you are a developer of a networked application, you should embrace encryption, no matter what you are sending. Only after a significant part of the traffic is encrypted will the Internet truly be an end to end network as it was originally intended. This is a good thing, and is the primary reason why the Internet has flourished to date.
Until then, more and more intelligence will be stuffed into the network, and it will offer no benefit at all to the users of that network. It only serves to further the special interests of large corporations and government, and will continue to be severely abused. It only serves to make the network more expensive, and one thing is for certain; it won't move the data any faster.
Only after this becomes a reality can we really concentrate on making the network faster and better, rather than inventing new ways to squeeze more money out of people for the same crappy infrastructure.
Henry Ford did build a car that not only used hemp in the construction but also ran on ethanol alcohol made from hemp. Before this Rudolph Deisel designed his deisel engine to run on hemp oil as well as other vegetable oils. In 1898 when he demonstrated his engine at the Paris Expo he had it running on peanut oil. A History of Biodiesel/Biofuels. In the 1930s a study by MIT found an acre of hemp would produce more paper than an acre of forest. Yet despite, actually as it treatened many wealthy and power people because of, the industrial advantages of hemp hemp was made illegal via the 1937 Marihuana Tax Act. Then as president Nixon had a study group to study whether hemp should be made legal again. However he said no matter what they concluded he would never agree to make it legal, which is what the study concluded.
FalconShould there be a Law?
Some people say the only acceptable advertising communications via e-mail is opt-in, everything else should be illegal. Call me crazy, but I'm kinda of the opinion that people should opt-in to spam blocking.
I realised something the other day that is a bit frightening to contemplate: some people actually like spam. They like junk mail in their snail mail box too. They like hearing about the new things they can buy and how much things cost. They like to hear about the things happening in their local area with "free" newspapers that are funded almost entirely by advertisements. Ultimately, they go and buy the products in these advertisements, which is why advertisers continue to advertise.
As long as these people continue to exist, is it even right to make spam illegal or, as many ISPs do, block it without even asking the receipiant if they want it blocked?
How we know is more important than what we know.
Much of what's in there is the classic telco dream - virtual circuits, charged by usage. What's being proposed is not the next Internet. It's the next ISDN.
Remember what went wrong with ISDN in the United States. The US telcos tried to use it as a way to get away from flat-rate pricing for local voice calls. That made it a non-starter for voice. The data pricing was so high it wasn't even feasible for data in the era of dial-up.
The Stanford "clean slate" document is basically "ISDN 2.0". Or, at the bulk level, "ATM 2.0".
From their own words, the agenda is clear - create a billable Internet where the price of each service can be cranked up by the service provider to the point that maximizes the provider's revenue.
There are times when I'm embarrassed that I graduated from Stanford computer science. This is one of them.
The idea of "updating" the internet makes me feel very much the same as when people talk about rewriting the US Constitution: we have a brilliantly conceived but outdated thing which could use an update to meet current circumstances impossible for the originators to have envisioned.
However, in the same vein, I'd be totally against it: I simply cannot see in the current world the ability to pull together an equally brilliant group of people who could do the task with an equal political objectivity. Indeed, as the internet is an acting infrastructure and not simply a set of rules on paper, it would be even more necessary to pull together resources from various who all have very different and conflicting biases. The BEST one could hope for would be something "designed by committee" ala the shuttle or the EU constitution. At worst, you're going to have interests conceding power in various facets to each other to suit their various needs. How would you like the internet *designed* by the RIAA? By the Republicans? By the Illuminati?
Thanks but no. I'll keep the creaky, leaky thing we've got. At least at it's CORE it's a fundamentally good thing. We just have to keep patching it.
-Styopa
At the risk of being modded down, I'd have to disagree. We've got to at least keep the current compromise between interests in law and order and interests in privacy (trying to be neutral here), otherwise we'll lose it. As it stands, if you get caught in a bittorent swarm of a pirated file, the **AA at least has a lead to you. We have no reliable automatic snooping system, and we have safety in numbers. I feel anonymous enough.
You know, there is a difference between trolling and pointing out the flaws in your reasoning. Just saying.
OK, here are a few of my "ideal world" wishes. Deciding their technical feasibility in real life is left as an exercise to the reader.
If you disagree, post your argument. (-1, Overrated) isn't your personal censorship tool for views you don't like.
Yup, some "needs" are just impossible to meet with the Internet in its present state. Like the "need" for a single agency to monitor all Internet traffic. Or the "need" for some folks to control every physical traffic channel. Or the burning need of one familiar industry group to be able to decide unilaterally which computers are "trustworthy" enough to connect to the Web.
Actually, we've long had other networking protocols that satisfied all these "needs". In fact, pretty much every network ever invented has satisfied them, except for the Internet Protocol.
The reason that IP won was that it's the only one that scales up to the size we have now. If you implement any of those "needs", you restrict your network to a small subset that doesn't violate that "need".
Organizations tend to prefer nice, neat setups that are organized hierarchically and can be monitored and audited. This is very useful for a single organization. But it isn't workable for a universal system. That requires parallel, independent development of the parts. If there's a central authority with local veto power, the system can't grow past what that authority's management can understand.
With any sort of central controlling authority, you can't have the explosion of development that has happened on the Internet. This can only happen if people have a way of developing what they want on their own. We can see this pretty clearly by comparing it the cell-phone system, which has the potential to give everyone full access everywhere and make the Internet look puny in comparison. But it's blocked by being limited to only devices and apps that the cell-phone companies' management approve and permit.
For a "new, improved Internet" to succeed, it must make independent local development easier than the current Internet. If it has any sort of controlling central authority, it will just remain a niche player that can't be adopted by enough people and expand to replace the current Internet.
Those who do study history are doomed to stand helplessly by while everyone else repeats it.
I'm sick and tired of waiting thirty seconds or more for somebody's slow ass Web server or puny pipe to feed me my porn!
This is nearly as bad as twenty or thirty years ago sitting at a green screen dumb terminal waiting for the mainframe to respond. At least then the wait times were shorter!
Not to mention the times the sites are totally down, or "you do not have permission to access this page" because some moron misconfigured his Apache Web server. (Remember that idiot in some Southern city who thought the site was hacked because the Apache configuration page was up instead of the home page?)
Run stats on your goddamn Web sites! Then buy another box or pay for more bandwidth! Or better yet, get the fuck off the Net because you don't know what you're doing!
Are you listening,
Anybody who thinks the Net is ready for "software as a Web service" is out of his goddamn mind. No company in its right mind would ever trust company business to the Net as the only option. It's hard enough to get the stuff on the company servers to work right. Trust somebody ELSE to do it right? It is to laugh,
Richard Steven Hack - This sig is TOO GODDAMN SHORT TO DO ANYTHING USEFUL WITH! MORONS!