IP Apocalypse - RIPE confirms available IPv4 addresses will run out in November

[vanished]

6 hours ago, LAwLz said:

Adding another octet or two would mean that we would probably encounter the same issue again in maybe 100 years or so.

IPv6 was designed to be future proof and as a result they really cranked up the number of available addresses.

Also, IPv4 has a lot of tings that needs improving, so they wanted to redesign the entire packet anyway, and at that point why not do it "properly" all the way?

IPv6 addresses are 4 times as large.

192.168.81.61.132.231.241.184.192.128.177.122.188.118.6.1 would have probably have been even harder to remember than the hexadecimal system IPv6 uses.

I was thinking about this just the other day actually.  Ignoring reserved ranges and just doing a simple calculation, I suppose IPv4 has 255^4 potential addresses and IPv6 has (16^4)^8, so that's about 80 octillion times more for a total of about 3.4 x 10^38, aka enough numbers to give 57000 unique IDs to every microgram of the entire planet Earth, assuming I haven't messed up that calculation.

 

I'm all for future proofing but I think they may have over done it lol

[The1Dickens]

So, don't know if this has ever been stated/assumed, but will all the IPv4 addresses become VOID when IPv6 gets a foothold? Or will IPv4 just be the "Old Internet" while "New Internet" will run off IPv6?

[dalekphalm]

18 minutes ago, The1Dickens said:

So, don't know if this has ever been stated/assumed, but will all the IPv4 addresses become VOID when IPv6 gets a foothold? Or will IPv4 just be the "Old Internet" while "New Internet" will run off IPv6?

I imagine that a lot of the internet will continue to run on v4 until there's a compelling reason for them to switch, while newer installs (or existing upgrades where they reach capacity and need more addresses anyway) will start to move over to v6.

 

That may mean that v4 addresses might start to free up in some capacity over time. Whether people will start to re-use them is anyone's guess - but given the vast number of people who want to hold out on v4, I'd guess that unless the regulatory authorities force people to adopt v6, some people will start to go after those old v4 addresses.

[leadeater]

On 11/1/2019 at 4:17 AM, LAwLz said:

especially in shared environments like AWS and Azure, NAT is a need since you have multiple tenants with overlapping IPs.

That's where SDN methodology comes in to save the scale issue, or really just throw more resources at it over a distributed scale. Pure network engineers and architects will likely thank you for it too as they no longer have to design a NAT solution at their network layer, "NAT all you like just don't make me do it or make it my problem".

 

It also allows hosters to let customers deploy their own security appliances like FortiGate VM or Citrix ADC, stuff like that.

 

Tenant isolation isn't really done with NAT though, it's done with packet encapsulation and tunneling, VXLAN for example. Providers will NAT tenant traffic exiting the tenant boundary if the resource isn't given a public IP but that NAT'ing doesn't really solve the overlapping IPs because that only addresses network border. Suspect that is all you were referring to though, just some blah blah etc for others if they are interested.

[BachChain]

On 10/31/2019 at 2:49 PM, Ryan_Vickers said:

I'm all for future proofing but I think they may have over done it lol

I have no doubt that they had the exact same thought back in 1981 when they were designing ipv4

[vanished]

2 minutes ago, BachChain said:

I have no doubt that they had the exact same thought back in 1981 when they were designing ipv4

Probably, but look at the example I gave.  It's physically impossible to ever come close to running out of addresses while this species is still primarily located on one planet

[LAwLz]

1 hour ago, Ryan_Vickers said:

Probably, but look at the example I gave.  It's physically impossible to ever come close to running out of addresses while this species is still primarily located on one planet

Well a lot of IPv6 addresses will be wasted.

The idea behind IPv6 is that end users, for example you and I, will be given an entire /48 block of addresses for use at home by their ISP. So each individual household will be able to create 16 million networks with 254 hosts each.

 

Overkill? Yes, very. But at the end of the day, it works. If you're redesigning the entire thing why not go all out and future proof the hell out of it?

[vanished]

3 minutes ago, LAwLz said:

Well a lot of IPv6 addresses will be wasted.

The idea behind IPv6 is that end users, for example you and I, will be given an entire /48 block of addresses for use at home by their ISP. So each individual household will be able to create 16 million networks with 254 hosts each.

 

Overkill? Yes, very. But at the end of the day, it works. If you're redesigning the entire thing why not go all out and future proof the hell out of it?

Wow that sounds kinda wasteful lol

But yeah, I don't mean to say they shouldn't have gone that far.  It's not like it really costs anything to may as well make sure there's plenty, I just mean to say that I'm extremely confident that there will in fact be plenty with this system

[jagdtigger]

2 hours ago, LAwLz said:

end users, for example you and I, will be given an entire /48 block of addresses

Knowing ISP's they gonna see a business opportunity. Since there is no NAT for IPv6 they gonna charge for every single extra IP above the the one included in your contract... (I already heard rumors that some ISP's starting to play around with the idea that they could charge ppl for public IPv4 addresses, while there is no option for Pv6.)

[justpoet]

On 10/28/2019 at 2:18 PM, jasonvp said:

My imbecilic ISP won't deploy v6 for business class users for some oddball reason.  Verizon has had the same IPv6 status page up for years, "It's coming... soon!" etc, etc, etc.  Any time I call them, I'm given the same response, "Um.... IPv... wha?"

 

If I could get on native v6 without tunneling to HE, I'd be on it like white on rice.  During busier times of the day (eg: late afternoon), HE's tunnel servers get walloped and I end up buffering way too much.  I appreciate what they've done (for free!) but they're not the proper solution.  Native v6 is.

 

 

A little further north of you, FIOS gets IPv6 dual stack automatically.

IPv6 still breaks a lot of services and game servers and the like.  Blizzard used it by default when it was available for a while, but it made so many more connection issues that they had to revert.

[BachChain]

39 minutes ago, jagdtigger said:

Knowing ISP's they gonna see a business opportunity. Since there is no NAT for IPv6 they gonna charge for every single extra IP above the the one included in your contract... (I already heard rumors that some ISP's starting to play around with the idea that they could charge ppl for public IPv4 addresses, while there is no option for Pv6.)

You can't really provide just one address. It has to be a range or else client configuration and routing becomes a nightmare

[jasonvp]

2 hours ago, justpoet said:

A little further north of you, FIOS gets IPv6 dual stack automatically.

I suspect they have it dual-stacked here for Resi, too.  Business class?  Nope.

 

[Colonel_Gerdauf]

5 hours ago, LAwLz said:

Well a lot of IPv6 addresses will be wasted.

The idea behind IPv6 is that end users, for example you and I, will be given an entire /48 block of addresses for use at home by their ISP. So each individual household will be able to create 16 million networks with 254 hosts each.

 

Overkill? Yes, very. But at the end of the day, it works. If you're redesigning the entire thing why not go all out and future proof the hell out of it?

That from what I can tell is very useful for large buildings under a single network infrastructure. While it might look absurdly overkill today, I can easily imagine business LANs exploding is client size thanks to the proliferation of IOT devices and ITTT.

 

For home networks though... I am not so sure. I have not been reading too much about the mathematics and logic behind IPv6 addresses, but I assume that there is a designation of private addresses similar in size and subnetting capabilities to 192.168.[SN].#.

 

Now that I am thinking about it, instead of retiring IPv4 entirely, would it make more sense to still allow it in LANs? That way, you would only have 1 or maybe 2 IPV6 addresses to deal with, connecting to the ISPs, and have however many IPV4 LAN addresses that they need to use, up to 4 billion unique addresses on up to 65K separate subnets. If there is a business that needs more than 4 billion addresses on a singular local point, I would be very surprised.

[BachChain]

13 minutes ago, Colonel_Gerdauf said:

For home networks though... I am not so sure. I have not been reading too much about the mathematics and logic behind IPv6 addresses, but I assume that there is a designation of private addresses similar in size and subnetting capabilities to 192.168.[SN].#.

The point of IPv6 is that every device has at least one globally routable address. The idea of private addresses and Nat like ipv4 is explicitly not supported as part of the IPv6 spec.

[justpoet]

You can think of it as overkill, but then you realize that with IoT trash, every shade, every thermostat, every smoke detector, every vent, every light, every switch, every speaker, every appliance, every phone, every computer, every energy/resource monitor, etc…all are also an IP.  It starts to add up really quick in even a medium sized building in the envisioned "smart" future.

[leadeater]

1 hour ago, Colonel_Gerdauf said:

and ITTT

What does that start for.

 

Also argh yet another stupid acronym 

[vanished]

1 hour ago, BachChain said:

The point of IPv6 is that every device has at least one globally routable address. The idea of private addresses and Nat like ipv4 is explicitly not supported as part of the IPv6 spec.

From what I could tell it certainly seemed to work that way, but good to know this is true and by design

 

_{still don't like it tho}

[Donut417]

4 minutes ago, Ryan_Vickers said:

From what I could tell it certainly seemed to work that way, but good to know this is true and by design

 

_{still don't like it tho}

Well, the issue is we have what 4 billion ish IPv4 addresses and what like 7 billion living on Earth. The numbers just dont equal out. 

 

My question is when we got strictly to IPv6, id say in about 15 or 20+ years, at the rate we are going. As it took about 20 before ISP's even started deploying it. I wonder how home networking hardware will change. Id think we would still have some type of box between the modem and network, acting at the very least as a firewall. 

[vanished]

4 minutes ago, Donut417 said:

Well, the issue is we have what 4 billion ish IPv4 addresses and what like 7 billion living on Earth. The numbers just dont equal out. 

Yeah it's amazing it's lasted this long frankly and I have to assume it's only because of NAT that it works.  Every home gets one address even though they might have 4 computers and 4 phones, plus TVs, shoes, water bottles, and juicers all connected to the network as well.

Well, that and billions of people living without internet access.

Quote

I wonder how home networking hardware will change.

Well, I don't think any of it will need updating since even gear from many years ago already supports IPv6, but at the rate it's coming I'm sure most of the existing stuff will be replaced by then anyway.

Quote

Id think we would still have some type of box between the modem and network, acting at the very least as a firewall. 

I would absolutely expect and hope so yes

[leadeater]

5 minutes ago, Donut417 said:

I wonder how home networking hardware will change. Id think we would still have some type of box between the modem and network, acting at the very least as a firewall. 

There will still be a network device between public and private networks, like a firewall as you mentioned. You'll need a device like that for things like authentication with your ISP for the connection (PPPoE for example) and other possible functions like media conversion (coax to ethernet etc).

[Donut417]

1 minute ago, leadeater said:

vice like that for things like authentication with your ISP for the connection

That would be the modem I mentioned. Not every person uses a gateway. I refuse to. But you do make a valid point for PPPoE. Ive never used a connection that required it. 

[leadeater]

1 hour ago, Donut417 said:

That would be the modem I mentioned. Not every person uses a gateway. I refuse to. But you do make a valid point for PPPoE. Ive never used a connection that required it. 

No but if you end up with a fibre connection and now days it's most likely going to be a PON, typically GPON more rarely EPON, so a minimum of a ONT will be required (HFC/cable/dsl same deal as above). Then you need a router, could be yours or could be the ISP's but it doesn't remove the need for that. Even on an unauthenticated connection you still need a network device to connect to the ISP provided connection and establish the session either by DHCP or static. Just chucking a switch on the connection then plugging your PCs in will most likely get your connection cut off when your ISP detects crap coming from you that is not allowed over the internet like NetBIOS broadcasts or Bonjour or a myriad of other crap that is LAN local.

 

IPv6 doesn't remove the requirement or change the design of public and private networks in that regard to network boarders and gateways, they are still required. There will always be something between you(your computers/devices) and the internet, only IP addressing changes.

 

Basically yes everyone does use a gateway, there isn't a possible way to connect to the internet without it. These gateway devices can, and in most countries are multi role devices that do everything (modem/router/firewall/etc).

[jagdtigger]

8 hours ago, BachChain said:

You can't really provide just one address. It has to be a range or else client configuration and routing becomes a nightmare

Its not that hard. Every node before the consumers assigned a block( a /112) and  each consumer gets a single address by default from that block

[Tellos]

Thankfully my company implemented V6 years ago and only uses v4 for specific things.