What lies ahead for the Internet, WWW and TCP/IP

[guestgulkan]

I believe these three(the Internet,

http://WWW and

TCP/IP) are rapidly becoming victims of their own success.
OK, Internet & TCP/IP didn't take the world by storm on their debut, but we are all aware of the limitation of the IP4 addressing method.

With sound and video going over the net - a standard IP packet of 1500 bytes seems really restrictive to me, all that header overhead and TCP acking/nacking.

Is IP V6 going to see the light of day - or will it become overtaken by events?

The

http://WWW has

rapidly becoming clogged with unused/stale/defunct/etc.. websites/pages and blind links.
It seems doomed to drown in it's own waste!

 

[sleipnir214]

Sorry, no productive comment on IPv6, but here's an April Fool's Day, 1994, RFC which takes an historical look at IPv9:

http://www.faqs.org/rfcs/rfc1606.html

I don't think I agree with you that the world wide web is becoming clogged with its own effluvium. While I grant you that there are an enormous number of defunct sites, I find new informative sites every time I perform online research. And it often seems to be the case that that the defunct sites were set up by people who have run out of things to say or who got frustrated at the learning curve of HTML and quit doing it.

http://WWW is

like a living tree -- limbs fall off and new ones grow in their places.



Want the best answers? Ask the best questions:

http://www.catb.org/~esr/faqs/smart-questions.html

TANSTAAFL!

 

[Tarwn]

Sleipner, thanks for an RFC I hadn't caught in the past, though I still think the Memo from the Consortium for Slow Commotion Research is my favorite.

As to IPv6, as with all market overhauls, as late as possible...how long did it take for businesses to finally start doing something about Y2K? The general concept is don't spend a dollar today you can put off until tomorrow...or better yet make someone else spend in order to come up with yet another way to save you your original dollar.

-Tarwn

_{01010100 01101001 01100101 01110010 01101110 01101111 01101011 00101110 01100011 01101111 01101101}
^{29 3K 10 3D 3L 3J 3K 10 32 35 10 3E 39 33 35 10 3K 3F 10 38 31 3M 35 10 36 3I 35 35 10 3K 39 3D 35 10 1Q 19}
Get better results for your questions: faq333-2924
Frequently Asked ASP Questions: faq333-3048

 

[sleipnir214]

<aside>
I think my all-time favorite is RFC 1149, &quot;A Standard for the Transmission of IP Datagrams on Avian Carriers&quot;, released April 1, 1990. But only because someone else actually amended it with RFC 2549 on April 1, 1999 and because a group of crazy Norwegian geeks actually implemented it. The actually used carrier pidgeons to send pings back and forth across their &quot;network&quot; (

http://news.com.com/2100-1001-257064.html?legacy=cnet).

</aside>

I think a big stumbling block to IPv6 is legacy IP stacks.

As I understand it, if your network is IPv6, then IPv4 networks are available to you. But I don't think the converse is true.

That means groups with an interest in VPN WANS would find IPv6 attractive -- they get the latest networking, and are still able to get to the outside world. But an organization like amazon.com can't go IPv6, except as a second IP stack -- there's not enough customers using IPv6 to make the upgrade economically justifiable.

Want the best answers? Ask the best questions:

http://www.catb.org/~esr/faqs/smart-questions.html

TANSTAAFL!

 

[dilettante]

I think the &quot;1500 byte&quot; thing is an Ethernet limitation/convention, nothing to do with IP per se.

 

[chiph]

Whilst there are still many addresses unallocated the available space has been sharply decreased. The discovery of intelligent life on other solar systems with the parallel discovery of a faster-than-light transport stack is the main cause. This enables real time communication with them, and has made the allocation of world-size address spaces necessary, at the level 3 routing hierarchy. There is still only 1 global (spatial) level 2 galaxy wide network required for this galaxy, although the establishment of permanent space stations in deep space may start to exhaust this. This allows level 1 to be used for inter-galaxy routing. The most pressing problem now is the case of parallel universes. Of course there is the danger of assuming that there is no higher extrapolation than parallel universes...

Actually, I've been doing some reading recently on the development of the inter-planetary IP protocol (

http://www.ipnsig.org/).

Because of the light-speed delays, the network would function more as a store & forward system, rather than a real-time streaming system (no live video chat from Mars!).

The amusing thing is that the DNS top-level domains would be extended to include the solar system and planet you're on. So Tek-Tips would be:

http://www.tek-tips.com.earth.sol

and I imagine they could extend this to include your quadrant and galaxy. Right about that time we'll be ready to go from IPv6 to IPv9!

Chip H.

 

[AIXSPadmin]

After going to the IPN site (ipnsig.org) I was surprised to learn of DARPA and NASA being involved in the studies.

 

[spewn]

on this topic, i believe the 'web' will correct itself...although we have everyone and their brother building web pages, the leaves will be shaken from the tree if they can't improvise, adapt and overcome.

anyone can build a webpage; not everyone can build a web business, though. and even less than that can turn that business into an empire.

Q: where is the internet going to be in the years ahead?
A: wherever i need/want it to be.

and with that, i'm out.

- g

 

[manarth]

thanks to chiph for the ET post - I have been considering for a good while now the implications of extra-Sol-communications... so it's nice to see someone ahead of me!

As for the web; I agree with Spewn - I think it's self correcting... basically, a dodgy website has to paid for either way; if it's that bad, it's not going to feature on Google (do the other web search engines still exist? I know not!) Even if it remain running for eternity, the impact to the web is a. the entry in the DNS lookup may slow a DNS search by a bit and b. it's given an IP address which could be given to some(one/thing) else.
Either way, the owner has to pay for it, so the inevitability of the cost-benefit transactions in a capitslist world should win out. hopefully =)

At the end of the day, webhosting costs money...and yes, there'll always be bad websites.

Re: running on IP4 trying to access IP6: of course it ain't ever gonna work. If your car only does 80kmh, it cannot drive at 100kmh: the only solution would be to backward compatible the entire IP system, which would - by default - limit the entire system to IP4 anyway (80 kmh) - and so render any prospect of inprovement null and void.
That said, there's no harm in allowing IP6 systems to access IP4 networks.


<marc> i wonder what will happen if i press this...[ul][li]please give feedback on what works / what doesn't[/li][li]need some help? how to get a better answer: faq581-3339[/li][/ul]

 

[Grenage]

Would you want anything bigger than 1500 ? I mean wouldn't it just quadtrouple the chances of it becoming corrupt if you had 5k packets.

 

[dilettante]

One big problem people run into is MTU-mismatch. Data sent through ethernet and then through FDDI used to get fragemented badly because ethernet uses 1500 while FDDI used 4352. Smarter &quot;coupling devices&quot; have helped resolve this sort of problem to a large extent, but this was why ethernet over fibre was so popular for awhile: FDDI had better performance, but in real networks the fragementation seemed to put hybrid networks at a disadvantage.

You have to remember that when ethernet was born a lot of it was used for Telnet. This meant there were lots of packets carrying one-byte payloads. Memory was also expensive, so huge buffers in endpoint adapters came at a premium.

1500 was a sort of compromise, and minimizes the overhead of single-bit failure retries.

Many other approaches and MTU sizes have been experimented with over ethernet technologies. My guess is that no compelling advantage was ever found that warranted obsoleting existing hardware, firmware, or software.

People still work at this though, and some gigabit ethernet setups use MTUs giving payloads of 8K or more (9000 bytes is the max MTU many gig cards can be set to). Even so, most networks still use 1500/1504 over gigabit ethernet. This is probably due to much legacy firmware optimization for 1500 that will be with us for a long time.

It is pretty well entrenched.