No More IPv4 - IPv6 is the Future

[DemoBytom]

IPv4 being officially out of addresses won't do anything to the internet, given how prevalent IPv6 is already.


There are enough addresses that if you cut up the earth into pieces of equal mass, where each piece was the mass of a paperclip, each of those pieces would have more unique addresses available to it than all of the addresses in IPv4.

There are 2^120 or 1.33*10^38 IPv6 addresses. There are 1.33*10^34 atoms on the surface of the Earth. So we could assign an unique IPv6 address to every atom on the surface of Earth and still be left with enough to do another 100 Earths..

http://www.edn.com/electronics-blogs...Head-of-a-Pin-

[wowaccounttom]

this will be a nightmare

that is all.

[zealo]

Calcing subnets in ipv6 though

That's definitely a downside to it in terms of ease of use, but can live with it.

[Nadiru]

IPv8 soon. Probably 15 years. Internet of Things will eat up so many addresses you'll never believe it.

[unbound]

If you haven't heard, IPv4 addresses have officially been depleted. The future (it's been around for a bit now) is IPv6. Do you think IPv4 addresses running out will have any impact on the way the internet grows over the next year, five years, ten years? IPv6 is said to have an almost limitless supply of addresses.

Been shifting to IPv6 for several years now (up to about 15% of the internet traffic in the US now...Belgium leads at about 33% of their traffic being IPv6). We'll likely be hybrid (both IPv4 and IPv6) for a couple of decades before IPv4 is finally fully abandoned.

There are headaches associated with the changes, but it should be extremely rare to affect any end user.

- - - Updated - - -

IPv8 soon. Probably 15 years. Internet of Things will eat up so many addresses you'll never believe it.

Not anywhere near that soon.

IPv4 maxes out at about 4,300,000,000 addresses (which got extended using creative methods like proxy servers).

IPv6 maxes out at about 340,000,000,000,000,000,000,000,000,000,000,000,000 addresses. It'll take a long time before we run out of those addresses even with exponential growth.

[Anjerith]

Fun fact; many many many devices are hard-built to only look at addresses matching IPv4 standard, and they are not magically patchable. So on the off-chance that the industry ever DOES just decide to drop IPv4 for IPv6 in the next 10-ish years, expect a LOT of problems on a LOT of random Wireless (especially) devices.

[unbound]

Fun fact; many many many devices are hard-built to only look at addresses matching IPv4 standard, and they are not magically patchable. So on the off-chance that the industry ever DOES just decide to drop IPv4 for IPv6 in the next 10-ish years, expect a LOT of problems on a LOT of random Wireless (especially) devices.

Yep. That's why we'll be in a hybrid IPv4 / IPv6 world for a couple of decades.

[zealo]

IPv8 soon. Probably 15 years. Internet of Things will eat up so many addresses you'll never believe it.

IPv6 offers so many possible addresses you could give every single internet capable device there is a static IP if rolling with the assumption their all capable with it.

On that note, It would be kinda neat if more ISPs started offering static IPs in IPv6 so you can run things like home servers on home networks more reliably without needing to jump through hoops to access them from the outside.

[DemoBytom]

IPv8 soon. Probably 15 years. Internet of Things will eat up so many addresses you'll never believe it.

Just 2 posts above.. you could assign an IPv6 address to every atom of the surface of Earth and be left with enough free addresses for another 100 Earths... We're not gonna run out, unless someone starts reserving ludicrous amounts of addresses (and by that I mean several magnitudes worth of IPv4-size ranges for 1 home network)..

[Masark]

Why? My guess, is because IP4 is more than capable to handle what we need... 10.x.x.x and 192.168.x.x is used in combination with NAT to handle what IP6 was supposed to do, and it's doing a great job...

No, it isn't "doing a great job". It's a shambling patchwork that was only ever intended as a stopgap to allow for the implementation of IPv6.

It breaks end-to-end connectivity, hampers security, and forces stupid hacks like holepunching.

The only reason IPv4 "ran out" is because when it was originally subnetted they allocated WAAAAY too many addresses to Class A in chunks that were FAR too large and so now the media conglomerates who got first dibs at the Class A space get to lease it out and line their pockets, hurrah.

Bullshit. Even if you reclaimed all the A blocks, it would buy you a year, at most. Likely considerably less than that. And then you're out again.

[imabanana]

this will be a nightmare

that is all.

Why?

One of the major French provider (Free) rolled it on all their network in only 5 weeks, and that was back in 2007. I've used it for years without having any issues, and lots of people have also been using it totally seamlessly for close to a decade now.

According to a study published by Google at the RIPE meeting in October 2008, Free is probably the largest native IPv6 ISP in the World.[45] By end 2008, almost all French IPv6 traffic measured in the study comes from Free customers.[45]

Free deployed the IPv6 infrastructure in only 5 weeks, from 7 November to 11 December 2007, thanks to an innovative 6rd (IPv6 rapid deployment) proposal by Rémi Després.[44]

[Planeshaper]

There are 2^120 or 1.33*10^38 IPv6 addresses. There are 1.33*10^34 atoms on the surface of the Earth. So we could assign an unique IPv6 address to every atom on the surface of Earth and still be left with enough to do another 100 Earths..

http://www.edn.com/electronics-blogs...Head-of-a-Pin-

IPv6 address space is 2^128 though, not 2^120.

But I like this game. So, let's continue!

The earth has a mass of 5.972*10^27g. At most, 25% of earth's mass is Silicon, in some form of silicate. The mass of Silicon is 28.0855g/mol, and 1 mole is 6.022*10^23 atoms. It requires approximately 10^6 atoms of Silicon to store a single bit. There are 2^128 addresses in IPv6. There are 8 bits in a byte.

There is an IPv6 address for every 12 kilobytes that can be stored by earth using our current method of magnetic storage.

[ReliableSite]

The fun fact of the day is that ISPs hadn't even bothered to make their equipment compatible with IPv6 till shit started to hit the fan (ie. till it was too late), and home routers still need to be compatible with it in order to connect to said ISP, though I believe most are. Plus, isn't IPv6 supposed to be more secure and faster than IPv4?

Yes, IPv6 is both more secure and faster.

- - - Updated - - -

"No, it isn't "doing a great job". It's a shambling patchwork that was only ever intended as a stopgap to allow for the implementation of IPv6.
It breaks end-to-end connectivity, hampers security, and forces stupid hacks like holepunching."

Agreed. There are a number of differences between the two and security is definitely one of them.

[Samurai]

The only reason IPv4 "ran out" is because when it was originally subnetted they allocated WAAAAY too many addresses to Class A in chunks that were FAR too large and so now the media conglomerates who got first dibs at the Class A space get to lease it out and line their pockets, hurrah. But with port forwarding each existing IP address can be multiplied 64k times and that's why there has really been no rush to conform to the IPv6 standard.

You need a unique IP if a address is using an SSL Cert.
Which with the changes Mozilla and Google are doing with their browsers, will make more sites get a Certificate.

[Deleted]

There are 2^120 or 1.33*10^38 IPv6 addresses. There are 1.33*10^34 atoms on the surface of the Earth. So we could assign an unique IPv6 address to every atom on the surface of Earth and still be left with enough to do another 100 Earths..

http://www.edn.com/electronics-blogs...Head-of-a-Pin-

Actually an IPv6 'address' is split in half between the network address and the host address. So there are actually only 2^64 unique unicast addresses. If you move to a different network then the network part will change but the host part (lower 64-bits) will remain the same.

Oddly enough there are actually *way* more multicast addresses than there are unicast addresses...

I went to my first Cisco "Year of IPv6" event around about 1999 .

It has been proceeding slowly - mostly due to the widespread use of NAT to extend the life of IPv4.

- - - Updated - - -

You need a unique IP if a address is using an SSL Cert.
Which with the changes Mozilla and Google are doing with their browsers, will make more sites get a Certificate.

Actually sites tend to tie their certificates to domain names rather than IP addresses - the addresses don't need to be unique.

In fact you can run multiple web servers with different SSL certs off the same IP address.

[Masark]

You need a unique IP if a address is using an SSL Cert.
Which with the changes Mozilla and Google are doing with their browsers, will make more sites get a Certificate.

Not necessarily. Server name indication allows for multiple servers behind a single IP, though there's some old crap still around that doesn't support it.

[Deleted]

On that note, It would be kinda neat if more ISPs started offering static IPs in IPv6 so you can run things like home servers on home networks more reliably without needing to jump through hoops to access them from the outside.

Static IP is awful, if someone gets pissed at you and decides to DDOS you, you can't change your ip.

Also, dynamic dns servers works good for home servers, some routers even having built-in ones.

[Septik]

When I was in the IT business, back in the '90s, everyone said "Within 2 years, we'll switch to IP6"..
In the '00s, everyone said "We'll switch to IP6 within 5 years"..
Here in the '10s, IP6 talk have been virtually non-existant..

Why? My guess, is because IP4 is more than capable to handle what we need... 10.x.x.x and 192.168.x.x is used in combination with NAT to handle what IP6 was supposed to do, and it's doing a great job...

thats funny cause i recall hearing the same way back in 1997 or so and then around 2005 i thought, wheres the new ip system?? and here we are today.

are we finally at that point where we have to switch, or is this just more of the same bs from the past twenty plus years?

[Deleted]

Yes, IPv6 is both more secure and faster.

- - - Updated - - -

"No, it isn't "doing a great job". It's a shambling patchwork that was only ever intended as a stopgap to allow for the implementation of IPv6.
It breaks end-to-end connectivity, hampers security, and forces stupid hacks like holepunching."

Agreed. There are a number of differences between the two and security is definitely one of them.

People say this but the only justification I know of is that IPsec is a mandatory part of IPv6 but is optional with IPv4. In fact few computers use IPsec end to end and web traffic is generally protected with TLS, which is the same for IPv4 or IPv6. There is also scope for more security in the neighbour discovery protocol but I doubt this adds much over WPA2, and isn't really the type of security most people are bothered about.

IPv6 has some features that allow for more hierarchical network numbering in the core but there are other initiatives (eg Cisco LISP) to manage that routing complexity.

[ReliableSite]

Static IP is awful, if someone gets pissed at you and decides to DDOS you, you can't change your ip.

Also, dynamic dns servers works good for home servers, some routers even having built-in ones.

True. All about that DDoS protection

- - - Updated - - -

I think that we can all agree that even though IPv4 has run out, it will be quite some time until IPv6 is used exclusively. Although there are benefits to IPv6, IPv4 still gets the job done.

- - - Updated - - -

I wrote a post about the advantages of IPv6 Vs. IPv4 that outlines some of the benefits a few months ago. Worth a read for those who don't know the difference.