Your next IP address in 128 bit
 

A 128 bit IP address (8 * 16 bit numbers)

17321.61258.6211.12158.57121.42118.7101.10719

I look forward to IPV6 actually getting implemented. I don't look forward to upgrading my servers and having stuff break...

Comments?

it will be sweet. so much clean high ways of internets :) i heard all ips will be gone in next 12 months, so Ipv6 here you come

i heard that too

My buddy who is a high up at Comcast is knee deep in IPv6 right now.

I have a suspicion that super access to grids of IPS will open new possibilities.

I remember when a programmer "talked" with the video card via 5 bytes: 2xpos, 2ypos and 1 read/write. In order to set a pixel, you would set the xy position and then read or write to the port.

Now, you have the entire video memory in 24 bits, in a grid. You can access any area you want. So much easier.

I can imagine where you can access parts of your website by IP, and deliver content, based on grids of data, rather than a single url portal.

Just a random thought though...doesn't really make better porn though :-)

Not a problem. Encapsulation fixes that for you. Just keep surfing.

i will start spamming again... think of all the free ip space available thats not blacklisted... :thumbsup



.

http://labs.ripe.net/Members/blazque...spam-over-ipv6

That would be a bad way to invest your time.

I have heard about IPV6 for 12 years since i was a server admin back in the day.
Wonder if it will ever really happen lol

Windows 2008 server or Windows 7 really happened :upsidedow

nothing, breaks, we run ipv4 and ipv6 together and have servers running both. Nothing breaks it all works just fine. Server wise.

However tradescripts/etc stuff thats tracks ipv4 address space is going to be a whole different issue at hand.

You are making little sense willis :Oh crap

You were wondering if it would ever really happen, and I tried to tell you it already have, through Windows 2008 server and Windows 7 just to mention a few.

:1orglaugh that was the funniest thing I ever read... of course spam is low coming form ipv6 ip's.. most spam today is coming from infected pcs that are using ipv4...

wait till ipv6 is readily available.. you'll see.. :2 cents:



.

Did you see the article had spam in the comments? :Graucho

Unless they find a way of spoofing the IP, the spam wont be delivered because non-existent email addresses within the ripe.net domain, will simply be refuced delivery. They might find temporary weaknesses or workarounds, but I belive IPv6 will be a hard hit in the face of spammers.

Mail services such as yahoo, google and MS will face immense challenges, because they will need to find a way to handle the massive Greylisting hitting them.

We converted our network over to iPV6 about 6 months ago as Netflix and others required it. Now that its done, we are clearly ahead of the market and it has been influential in many of our new contracts.

That being said, 90% of the people on GFY have no idea what we are talking about. :1orglaugh

Companies like you at Limelight must love the traffic shaping and QOS capabilities across your whole network.

wtf are you talking about? what does ipv6 have to do with ts and qos?




.

No offence, but I do think you need to educate yourself on the topic, if you dont see how it's relevant. You speak of IPv6 as if you think its address spacing only.

19 replies and no one has pointed out that ipv6 uses a colon instead of a dot? :1orglaugh:1orglaugh

ipv4: 192.168.1.1
ipv6: 2001:252:0:1::2008:6

I sale short Ip....

Stock up on IPs.

i understand ipv6.. and yes i have heard people say that it helps from the standpoint of security due to ipsec, etc..

but, is that based more on issues with nat than anything else..

also, i believe that from a hardware standpoint, the technology has mitigated a lot of the issues of the past with complex routing tables, etc..

do i believe that ipv6 when fully implemented will make the internet better.. i dont know..
a network is always going to be limited by its weakest point and i think that network capacity outweighs ipv6. its easy to design something on paper, but until its fully tested in a live environment, i wouldnt speculate on the outcome..

on paper, is ipv6 better than ipv4.. yes it is..

will it make a difference on the overall internet experience? time will tell.. :2 cents:




.

How do you specify and additional port in IPv6?

by utilizing [2001:252:0:1::2008:6]:80

i.e

http://[2001:252:0:1::2008:6]:80

however with AAAA records you can still call things like a normal A record for the most part.

Looks like we dont completely disagree after all. To complete the speculation, I often hear people say that all the benefits we gain from when its fully implemented, will even out the traffic increase and content demand. I dont know if that is true, but time will tell as you said yourself.

The only thing I know is the huge benefits I have seen myself in live enviroments and during labtests. My sites and services are almost exclusively Windows based, so the native QoS is not only a shortcut, but also increase the potential capacity of the infrastructure.

The only thing ipv6 has done is make network engineers job a bit harder. Mainly because of the 128bit address space. Its simply larger than ipv4 address space.

Lets put it this way. cisco and juniper just recently released router gear that supports 1mil-4million prefixes.

Lets look at todays data http://bgp.potaroo.net/as2.0/bgp-active.html

There are 330k ipv4 address prefixes being announced across the net. Get a couple providers sending you full routes and your maxing out your shiny 10-20k route processor card. Sucks doesn't it?

So lets put this into perspective, a /24 is the smallest netblock that is allowed to be announced to the world. Everyone knows how big a /24 is. There are 172k /24's being announced independently of other network address space. Thats a LOT.

Now lets look at IPv6. Given that ARIN is allocating out /32's worth of IPv6 address space. Which is in essence every every single person the equivalent of 4billion IPs. Yes they are giving out the entire ipv4 routing table equivalent to everyone who wants it as the minimum.

Now the SMALLEST IPv6 address prefix that you can announce as the current "standard" is a /48. There are 64,000 /48's inside of a /32. So yes this means. it will take less than 5 networks announcing every single /48's individually to fill the same amount of router memory as todays IPv4 network space. This isn't even taking consideration of the amount of memory each route uses since ipv4 is 32bit and ipv6 is 128bit, so each route entery is going to take more memory.

So with this being said, the routers that power the backbones are going to have a hard time keeping up with the prefixes, hell they currently are having this problem. Let alone your servers firewall filtering stuff. You have not seen anything yet.

now you are splitting this into a debate of hardware vs b/w..

personally, i dont know about the benefits from a hardware (server) standpoint and frankly i didnt think about that aspect. so, if you see increased performance from a hardware (server) standpoint then i would say that is an aspect that factors into the equation quite a bit.

actually i often think about what plays more into internet performance, is it a bottle neck at the hardware level (routers, switches, servers, etc) for a lot of sites or is it network congestion.

if we are looking at this from a hardware vs b/w standpoint, i would say that when testing my servers and infrastructure, i find that i see more performance issues due to congestion and hardware capacity at the routing and switching level..

so, i guess my point was that until the entire core infrastructure, even down to the smallest host, will play more of a role than implementation of ipv6..

and even then, how many hosts/ip transit providers will continue to upgrade theri infrastructure to handle the increasing traffic load..




.

sounds like a good time to buy routing and switching technology stocks..

i didnt know this was such an issue.. i thought that the manufacturers had continued to improve performance when it came to doing more tasks with the same amount of hardware resources..

so, based on what your seeing, is it safe to assume that regardless of how ipv6 is the overall performance is going to be effected by the infrastructure providers upgrading their networks?

good post...



.

Juniper stock, yes. Cisco still has their heads up their asses and their latest and greatest "Nexus" router is a total flop. Everyone who touches it hates it and has given it back, broken IOS, poor features and limitations.

Juniper on the other hand is lightyears a head of ciscos technology in every aspect.

Cisco and juniper have always exchanged "leading" products over the years, cisco was great then juniper was better then cisco came out with something stable and works wonderful. Then juniper has knocked the ball out of the park and cisco is still playing with themselves in the dug out.

Been a while since I looked into ipv6 (should probably do more I guess) but I find it hard to believe that they'd be allocating single IPs to different entities and expect that the global net would be able to handle routing? Wasn't that just for testing purposes years ago when everyone was using tunnels?

My comment is that it's nice that I'm not going to upgrade my servers myself.
That's why managed hosting roxxorz.

What the fuck was wrong with regular dots?

FIGHT THE POWER!

People still use tunnels if their hardware cannot support native ipv6.

But to answer your question if everyone switched to ipv6 _today_ the infrastructure could not support it.

Typically if a router can support 1million ipv4 routes it can only support 256k ipv6 routes....... So yes. we are in trouble.

Me too. 5 years ago.

We're running out of IP's right? But why can't they use letters in an IP?

20A.196.4G.4P

etc.

And that is exactly why IPv6 will help, because the router dont have to work as much.

Thats basically what IPv6 is. Using hex inside of IPv4 mathematically doesn't work. Don't ask me to explain it because frankly, I don't care to know exactly the details of conversions etc. :)

We've been running IPv6 in test mode for about 6 months. We do IPv6 routing with almost all of our providers. I'm also on Comcast's IPv6 test network at home.

ARIN allocated a small /32 to us...which only gave us 4,294,967,296 IP addresses! I hope we don't run out :1orglaugh

-- Bill

I can't wait either, I have a lot of plans with it already..

Each "dot" is 8 bits, so it can only have the value 0-255. (Or as a 32 bit integer, 0-4,294,967,296)

x.x.x.x is just the way it's expressed to humans in decimal form; you cannot fit any more data in since natively it's a fixed size integer.

It gets more complicated because the legacy class A, B, C, D system reserves huge chunks of address space back when no one ever expected that we'd have millions of hosts on the internet.

IPv6 just adds more bits on, and expressing it in hex makes it a bit shorter and easier to understand. You could display it as x.x.x.x.x.x.x.x.x.x.x.x.x.x.x.x if you really wanted to :)

One positive thing with the IPv4 situation is that allocations can be recycled - legacy allocations from days gone by, like a medium size company with a huge A class allocation of 16,777,216 IPs, can renumber into a smaller allocation and release those IPs back into the pool. Unused allocations can also be reclaimed. It's still a big headache.

we should go back to the binary system :warning

Sounds like a lot of geekiness going on in this thread.