Thread 106906949

It's been 3 decades

>>106906949 (OP)
IPv6 would have much higher adoption rates if it was just IPv4 but with 128-bit addresses. NAT is a much more sensible method of separating local networks from wider ones, rather than just assigning almost half of the address space to each subscriber.

>>106907015
>NAT is a much more sensible method of separating local networks from wider ones,
It should be legal to punch people through the internet.

you left out the worst one, evpn-vxlan

>>106907319
Why should any internet services know exactly what device I'm using to access them at any one time?

>>106906949 (OP)
>>106907782
Is IPv6 supposed to supersede NAT? That sounds like less security since it bleeds more information from the intranet into the internet.

cgnat sucks ass as a user.

>>106907782
You have no idea what NAT does.

>>106906949 (OP)
You forgot DHCP. SLAAC is much better and does not need a centralised server.

>>106907015
>if it was just IPv4 but with 128-bit addresses.
It is. It's 4 times the size.
The way it works is actually way more simpler because the larger address size solves a lot problems on its own.
It's written in hex because otherwise it would be too long to write.
If you want you can write an IPv4 using IPv6 notation:
192.168.1.1 would be ::c0a8:101.
192 is c0
168 is a8
1 is 1
01 is 1
You can also do the reverse and use IPv4 notation on a IPv6.
One of 4chan's IPv6 is 2606:4700::6813:ea20 which would be 38.6.71.0.0.0.0.0.0.0.0.0.104.19.234.32 using IPv4 notation.
As you can see, hexadecimal is much more compact.
IPv4 notation on an IPv6 address is not officially supported so it is not being used anywhere.

>>106907782
>>106907805
IPv6 addresses are ephemeral by default. A device can have multiple addresses, an IPv6 usually lasts 1 to 5 hours if not statically assigned.
If an ISP is not retarded it will give you at least a /64 subnet, which means the first half will be decided by your ISP and the second half is decided by you, your router or your devices, according to your configuration.
I have a /56 which means my ISP controls the first 56 bits of the address and I control the remaining 72 bits.

>>106907015
>nooo, you can't just have a generic firewall with closed-by-default ports and you enable what you need
>you need some obscure IP sharing bullshit instead that got hacked on top of existing things

>>106907842
yeah. because one person is racist they blame the whole country.

>>106907805
NAT is no security feature.

>>106906949 (OP)
>port forwarding
I see no difference between forwarding port on IPv4 using NAT and opening port for particular IPv6 using firewall

>>106907782
The IPv6 device can assign any random addresses within your subnet, and in my network settings there is a privacy checkbox for this.
If you run a server, you can have your static one, if you don't, you can do any random thing within your /64.
If you want to run five webservers on different IPs on the same machine, all on port 80 and 443, you can do that as well.

I once evaded a SYN flood that was so powerful that the ISP blackholed my IP.... by simply using a different v6 IP within the subnet.
IPv6 also protects me from shodan and other scanning bullshit. While they can scan the whole v4 range easily, good look in randomly guessing the v6 i am using.

>>106907873
NAT is a done at the router and takes away a ton of compute resources.
You do not "open ports" on IPv6, they are already open, you simply don't drop the incoming packets, usually the default behaviour on the firewall configuration is to drop them.

>>106907859
>>106907897
The hexadecimal is fine. But while a device could use a random IPv6 address, in reality it just gets set to the MAC address unless specifically configured otherwise.

>>106907954
>while a device could use a random IPv6 address, in reality it just gets set to the MAC address
Literally a skill issue.
Every device supports it, because it is in RFC 4941. If your device doesn't, then blame your OS for not fully supporting IPv6.

>>106907954
Yeah that was the retarded behaviour it had before. In 2007 the IETF published RFC 4941 which fixed that privacy hole by randomising the last bits assignable by your devices.

>>106906949 (OP)
Wake me up when ipv6 has human readable addresses. I don't give a rat fuck how much better ipv6 is, I can't remember addresses that complex.
>Just use DNS to point to things you tard
Why? I can easily remember the pattern we use for stuff internally in the company, and adding DNS to the mix is just another point of failure when a piece of equipment gets put into or taken out of service, so say nothing of how much of a shitfest DNS propagation itself is.

>>106907859
>If an ISP is not retarded it will give you at least a /64 subnet, which means the first half will be decided by your ISP and the second half is decided by you, your router or your devices, according to your configuration.
>non retarded ISP
Meanwhile in reality half the ISPs outside of korea and some parts of Europe don't even support IPv6. You literally can't GET an ipv6 address on any of the fiber companies in my state. It's IPv4 all the way.

Frontier is /allegedly/ rolling out IPv6 support the same way they have been for the last 3 years. They were initially planning on something like /120, but their engineering team bitched about it. 2ish years ago they were waiting on firmware fixes from Juniper. Who knows what the hold up is, but it's been at least 3 years at this point and no progress.

Ironically, comcast has zero problems with ipv6, although you may have to inquire about it depending on the region you're in.

>>106907941
>NAT is a done at the router and takes away a ton of compute resources.
We've had hardware optimized for this for multiple decades now. It's a solved problem.

>>106908207
>I can't remember addresses
https://www.geeksforgeeks.org/techtips/how-to-open-notepad-in-windows/

>>106908207
>Wake me up when ipv6 has human readable addresses
you are the problem with computer networking. We gave up the fucking world for "human readability". FUCK YOU
Consider tor onion addresses, looks like
>jitiji3j38udfjsniggerjsd848wosl.onion
We get TLS for free because the address is the encryption key. We don't need PKI or any such bullshit. We don't need DNS. But people like you would complain that you can't easily say it to your gay boyfriend, and we give up so much just so faggots like you can "say it".

>>106908233
>Copy 15 different addresses and pollute your clipboard reel when you could punch in 4 short numbers for each based off a pattern of what category of device it is and where it's located in the building.

By the time you find the first address you want in your massive text file, I've already pushed the new firmware blob to 5 of the 7 3d printers that need the hotfix to run the new polymer filament and am pulling logs from the CnC lathe to see if I can figure out why it crashed both of the #7 reaming tools overnight before my coffee finishes brewing.

If you let monkeys design your network, sure, you have to search for stuff. If you put actual forethought into assigning address spaces and subnets, it's all intuitive and can be banged out in under 3 seconds on a numpad.

>>106908310
>We don't need DNS
Yeah, I agree, it all does seem so superfluous when you've used a protocol that functions just fine without any of the "readability" fluff.

>>106908424
you know you can assign hostnames to things, right? Life doesn't have to be shitty, you make it that way.

>>106908436
No DNS outages
No DNS registrars pulling your name because you said the nigger word
No yearly registration fees
No scam domains and domain-squatting
Fuck DNS, Fuck IPv4, Fuck NAT, fuck jannies

>>106908603
bro, you just do http://hotname.local
this works with mDNS, which doesn't need any DNS server, it's multicast.

And if this doesn't work, because your device doesn't implement basic standards, you just do the thing on your router. Your router is running its own dns resolution, right? Then you have http://hostname.lan or whatever you have configured.
If your router has an outage, you don't have a connection anyway.

You people are living in the past.

>>106908643
i think i did reply to the wrong anon, but my point stands

>>106908643
I don't want to. I want to put in an address and go there. That's it. Copy and paste exists. Bookmarking exists.

>>106908310
>We gave up the fucking world for "human readability".
And yet people like you insisted that IPv6 was the way forward despite all the pleading to the contrary by people who actually had to work with stuff in the real world. Is IPv4 perfect? Fuck no. The limited space is obviously a problem, but what the retards designing the v6 spec failed to grasp is that not everything in the real world works the way people in the tech sphere think it should. Companies will jump through insane hoops to avoid rebasing their infrastructure because doing so is risky and risks are expensive and scary.

Should you be direct connecting to IP addresses for anything, let alone everything? Fuck no, but it's going to keep happening because nobody in their right mind wants to setup all the extra crap to connect to a printer. When you start designing stuff based on abstract ideals and not what operations technology is dealing with in the trenches, you get pushback. The fact that people keep being surprised by this should tell you how dumb spec designers are.

I can't tell you how many times I've had to explain to PBX people that it doesn't fucking matter what they think the best way to use a desk phone is, people who have been using landlines since the 80s expect the phone on their desk to behave like they did back then, and they will get very angry when it doesn't. They're the ones making shot calls. You cater to them, or they make a stink about it. That's how the world works.

Meanwhile in delusional spec design land they're phasing out long term SSL certs because that's "more secure". Sure, it is in principle more secure, but I can guarantee you that sometime before 2030 we'll be reading about someone dying in the mountains because some cell equipment with an expired cert couldn't process a 911 call.

The real world is full of laziness and legacy. If your spec fails to account for how people actually use things, it's shit and people won't adopt it by choice. End of.

>>106908702
>we've always used fax machines here, if it isn't a fax machine no one is gonna use it!

>>106908541
You actually can't assign hostnames to a lot of industrial equipment because it's locked down.

Again, it's nice in theory, but in the real world things suck. We have several large format printers that can print D/E sized paper, and I think only one of them allows you to assign a hostname. Should this be the case? Again, fuck no. But it is the case, and telling people working with this shit that it shouldn't be is about as useful as the people protesting whatever the current monthly trending topic is on college campuses.

Reminder even without muh infrastructure, muh laziness there would be still a large resist against ipv6. FAGMAN don't want to make it easier for people to decentralize plus ipv4 rent seeking is basically free money. Tech industry is all about rent seeking.

>>106908750
Have you worked in the medical field at all?

They literally do that. You can't send medical documents via email, but fax is absolutely A OK. It's ass backwards, but that's literally part of HIPAA.

>>106908643
Sure, you could abstract it away and put it on the router, but that's just one more thing that needs to be correctly updated. In a shop where you have dozens of machines with static IPs being added/removed every month, you're just adding points of failure. Believe me, I wish there was a sensible solution that applied to everything we have to deal with, but there just isn't. It was enough of a nightmare for my predecessor to even get the network organized reasonably well. Once that got setup and we stopped having outages the suits were on board with aggressive network upgrades, but getting over that first hurdle took 4 years of back and forth arguing.

Now imagine that bureaucracy in a company the size of Frontier, and you'll start to understand why IPv6 isn't making any progress. Even the people that genuinely want it have no idea how the fuck to even begin implementing it.

>>106908784
This. So many industries rely on half the world being behind CGNAT, it's fucking crazy what they took from us
>>106908863
Yes the medical field is a regulated shithole, they still manage to lose all our medical data on the regular anyway

>>106906949 (OP)
>2025
>I still need to pay my ISP a bit more, if I want a public static IP, even if it is IPv6.
I hate this world.

>>106907859
38.6.71::104.19.234.32
fixed
@4chan @rapeape pls accept pr for gorgeous looks

>>106908863
>Sure, you could abstract it away and put it on the router, but that's just one more thing that needs to be correctly updated. In a shop where you have dozens of machines with static IPs being added/removed every month
In that case, you shouldn't have static IPs at all, because configuring those IPs on some shitty mini LCD display is more effort than entering the router config, checking the MAC of what tries to get an address, and assigning it a fixed one with a hostname.
So you have all your configuration on one single device, the router.

>>106908925
>they still manage to lose all our medical data on the regular anyway

It's almost as if a lot of regulations and specs don't serve any practical purpose beyond liability shielding.

>>106907859
No, it isn't, they added a bunch of bs baguage on top and you don't know what you are talking about.

It's just one Google search...

>>106908207
>Wake me up when ipv6 has human readable addresses.
fe80::1 is shorter than both 192.168.1.1 and 10.0.0.1
>Just use DNS to point to things you tard
This but unironically.
>adding DNS to the mix is just another point of failure when a piece of equipment gets put into or taken out of service
Then use mDNS for internal addresses, no failure on that, you just need 2 nodes, no server required. Just use anon-pc.local, coworker-pc.local, hr.local, ceo.local
>They were initially planning on something like /120, but their engineering team bitched about it.
Those "engineers" thinks IPv6 is just IPv4 but bigger, so they are retarded, it's much more than that.
>We've had hardware optimized for this
No matter how optimised it is, it will always be slower and clog up resources, the current protocol has none of that problems.

>>106909020
Shouldn't it be 38.6.71..104.19.234.32?

>>106909488
Which part of my post are you referring to?

>>106909113
that reminded me of a case where some printer config window only accepted ip addrs but the printers did have a hostname and all you had to do was regedit the config and it would work without any static shit

I don't know how to use ipv6 or NAT and I host websites
>gulp
what do I NEED to know?

>>106906949 (OP)
IPv6 usage has literally increased from 30% to 45% since 2020

>>106908702
based IT anon
I work in IT and I spend all day cursing the little limpdick faggots sitting on their bouncy ball chairs in some office in silicon valley making 6 figures who are so disconnected with reality that they don't understand how people ACTUALLY use the software that they are programming. Boomers are not going to adapt easily.

What happened to IPv5 though?

>>106911843
If your server(s) have a public IPv6 address (you can check by typing "ifconfig" in the ssh session and look for a public IPv6 in the list) then just add AAAA records in the DNS zone with the IPv6 address, do the same as you did with the A records (which are for IPv4).
That's it, your servers supports IPv6 now.

>>106912007
Anon that fucker is the boomer demanding shit remain the same because new is scary and a pain.

>>106912034
Its the windows convention. The odd ipv* is shit.

IPv2 was woke coded and shit by default.

>>106906949 (OP)
The most horrendous triple NAT (ISP CGNAT, modem NAT + internal NAT) was still less of a hassle to maintain than it is to setup a complex (i.e. more than the typical enterprise tier™ three hosts and two subnets) IPv6 only or god forbid dual stack network. And if your ISP doesn't 100% adhere to the intended IPv6 spec, so not as it is written but the nebulous way some aids riddled monkey intended, it's fucking over anyways since at that point you will have to deal with IPv6 NAT or whatever the IPv6 to IPv6 prefix translation shit was called which is aids on top of Marburg on top of the airborne version of the fucking plague.
It does NOTHING better.
At best it introduces new issues IPv4 doesn't have not even with triple or quadruple NAT and at worst has extremely similar issues to IPv4 but with a million times more complexity.
IPv6 doesn't have a reason to exist, never had one and never will have one.
>b-bb--b-b-but muh intergalactic empire
At that point we will either still use IPv4 with 10^24 layers of NAT or will have moved on to an actual replacement for IPv4 which IPv6 will never be.
>inb4 delusional cope ala b-b-b-b-but it was never intended as a replacement anyways
Well too fucking bad that it sucks even worse as an addon than it does standalone.

>>106906949 (OP)
IPv6 is fucking cancer and I am doing everything I can to stop it in its tracks
t. infrastructure architect

>>106910392
>Which part of my post are you referring to?
The idea that it's just ipv4 but with more bits.

They overengineered (badly) it.

>>106906949 (OP)
With ipv6 you could give a unique ip to every atoms and particles in the known universe just because.

>>106912501
Exquisite bait.
>IPv6 only or god forbid dual stack network.
IPv6 only is still not feasible as there are still too many IPv4 only servers.
Dual stack is easy, I don't know what are you doing.
>IPv6 NAT or whatever the IPv6 to IPv6 prefix translation shit was called
Yeah that is cancer, ISP sysadmins managing IPv6 exactly the same as IPv4 is why it has such a bad reputation. Switching to another ISP for shit like this should be the normal behaviour.
One of the reasons IPv6 was created is to get rid of NAT and those fuckers brought it back.
>It does NOTHING better.
>IPv6 doesn't have a reason to exist
How about no port forwarding, no confusing internal and external addresses, no split horizon dns, no sni proxies, no hairpin routing, no nat, restore end-to-end connectivity as the IETF originally intended, a simpler configuration process.
>At best it introduces new issues IPv4 doesn't have
The only complex parts are the ones that try to make it backwards compatible with IPv4.
>it was never intended as a replacement anyways
It was always intended to replace IPv4, ever heard of Sunset4? It will come a day that IPv4 will be turned off. It will probably be in 20 or 30 years but eventually it will come.

>>106912753
I know that, I was describing just the notation there.
The packet header and its flags are completely different, the logic is completely different and way more efficient, the main goal was to restore end-to-end connectivity by addressing the biggest flaws of IPv4.

>>106912034
not adopted by public and was quickly superseded

i got a new ISP and now i only have ipv4, is that good or bad?

>>106908702
I hate tech boomers so much it's unreal.
Just die you devoid of any brain plasticity retard.

The original idea was that everything can have it’s own IP address.
This turned out to be a terrible security idea, and NAT solves the IP problem.
Reminds me of the “mobile code” ideas of the 80s and 90s… seem insane now.

>>106913076
> 20 or 30 years but
Uh huh. Doubt.
Routers still come with serial ports

My ISP still doesn't support IPv6 even though they provide gigabit fibre. Fortunately they give real IPv4 addresses instead of CGNAT so I can port forward.

>>106906949 (OP)
IPv6 is fine, IPv4 is just easier to deal with.

>>106906949 (OP)
They will never take ipv4 from me!!!!!

>>106907920
Your granny loves you

>>106906949 (OP)
3 decades because ipv6 sucks ass. Everyone knows that it does and are dragging ass getting it implemented.... Because..... It sucks!!!! It fucking sucks!!!

Not only do half the world's ISPs not use it, but the rest do not have it set up correctly. Whoever wrote the spec forgot that people actually have to be able to learn it.

sorry whitoids, you need to check your privilege and change your racist IPv4 internet technology to make room for the 3 billion chindians plus another 4 billion africans and the diverse cultural enrichment they'll bring to the global village.

>>106913391
my previous one had ipv6, it wasnt gigabit.
>real IPv4 addresses instead of CGNAT so I can port forward.
how do i know if mine is cgnat and whats the usecase of porting it forward?

CGNATted IPv4 + public IPv6 is the inevitable future.
You want your own private IPv4? 99$ a month plus tip.

>>106913570
interesting, back in the day it was only americans and euros, no wonder people are nostalgic i think most of these people were more educated from what we have now.

>>106913580
>how do i know if mine is cgnat
get your external ip address and traceroute it. if there's a different intermediate ip address then it's probably CGNAT
>use case for port forwarding
running local servers (vpn, web, minecraft whatever) that can be connected to via your external ip address, won't work if it's a CGNAT. Also you'll get hit with other people's 4chan and other website bans due to sharing an IP address.

>>106912385
On the contrary I'm the one pushing for pretty aggressive changes in actual operations. I've actually managed to get corporate to blackball a few companies that would not provide cnc control equipment that would have required hacky workarounds to get online. We quite literally spent an extra 350k on laser equipment for our additive metal machines because the company we wound up going with was willing (excited even) to provide linux support via flatpaks and didn't require the orchestration server to run inside a fucking XP VM with internet access.

The difference between me and the techtards writing specs is that I actually broke it down in terms of dollars and wrote it up in managerial corporate speak. I showed them that the more expensive solution would pay itself off in a few years because of a laundry list of reasons, and while their eyes glazed over when I said the security concerns of the cheaper option were significant, they listened to the rest. About a year later one of our competitors half a state away got hit by encryption malware because of that xp machine plus bad backups and had 2ish weeks of downtime that made them miss deadlines on several major projects. You damn well better believe I brought that up in a strategic meeting. Corporate types love having their decisions validated, and their eyes lit up when they learned about the possibility of pillaging a few high value clients.

Spec writers live in a bubble where everyone is compliant with what is best in theory. Corporate doesn't care about that. They care about dollars, and until you can put things in clear cut terms with hard dollar amounts on the table, business executives will not care about what you have to say. v6 offers nothing to my company beyond additional costs because a ton of equipment won't work with it and even the stuff that does has any one of half a dozen different annoyances making it painful to work with, so any conversation about it is an automatic dead end.

>>106906949 (OP)
Oh shit a network thread, based. All of these technologies still rest on the disgusting abomination that is broadcast networking. Networking as a technology needs an overhaul as low as layer 2.

>>106913651
>Also you'll get hit with other people's 4chan and other website bans due to sharing an IP address.
>TFW you got your entire dorm banned from 4chan for 3 months that overlapped finals week for calling moot a faggot in a QA thread back in the day.

I probably helped 30 people pass that semester.

I don't know shit about networks. All i know is that i can't change my IP and it got banned on my favorite chan because someone else broke the rules so not i can't post there and have to wait weeks for it to change on its own

>>106914363
>have to wait weeks for it to change on its own
if you turn your router off and on you should get a new ip

>>106914440
nope

>>106914363
Your DHCP ip address is assigned to you by your ISP. So a release/renew won't work. You have to use a vpn to get around that.

>>106913570
Chinks have their own internet so they're mostly contained, it's jeets (and jeet-like countries), nigs (mainly nigeria) and SEAnigs that are the problem

>>106913656
Based nonbroadcast medium believer. Fuck ethernet, retard grade protocol.

T. CCIE SP, Nokia SRA

>>106906949 (OP)
You could probably eliminate the need for IPv6 by simply banning chinks, niggers, and jeets from the internet.

>>106907920
>>106908012
corporate policy can demand that your MAC be used for IPV6 address resolution

>>106906949 (OP)
>PCP
Correct! If you are wildly flailing on the ground fighting imaginary demons with your enormous hammer dick, the shoddy state of your network stack doesn't matter!

>>106914363
>>106914713
I switched to smaller local ISP for Gig fiber, and it came pre-banned lol. The whole ISP is ranged banned, and websties think Im behind a VPN

>>106908207
>human readable addresses
Yeah but FFFE is kino
And you never have to subnet ever again

>>106913570
Just use your own 0.0.0.0 0.0.0.0 network brown anon.

>>106911908
that didn't happen because of networks switching, but new (((devices))) at amazon and co.

>>106915513
I mean why not. CGNAT kind of is like this already. USANET and CHINANET using the same address range, which they overload onto whatever pool of global public IPs they're allocated. Otherwise they can just route internally which would probably be faster

>>106906949 (OP)
ipv6 addresses are a pain in the ass to remember. I know my home ipv4 by heart and can connect to my home server anytime I need

>>106912501
>The most horrendous triple NAT (ISP CGNAT, modem NAT + internal NAT) was still less of a hassle to maintain
No it's not, you are just dumb. You got uses to something overcomplicated and inefficient, because the ISP gave you a black box router who you just trust blindly.
You can't change anything, because you don't understand it and are unwilling to learn.
So you are fearful of touching it.
But you are in denial about this, so you make up reasons in your head about why your decision is totally reasonable and not just based on limited knowledge.

IPv6 unironically just works. While the configuration your ISP gave you is a monster.

>>106906949 (OP)
Not using Hex, fuck off.

>>106907458
retarded code monkey. EVPN is not used to free up IP space, it's used to band-aid L2 extension.

>>106906949 (OP)
IPv6 is unironically more complex than just using NAT (which with 8 bits of src ip (24 bitmask), 32 bits of dst ip, and 16 bits of src ports, and 16 bits of dst ports, effectively converts 32bit IPv4 into a 72bit address space). Also SLAAC was and still is retarded.

>>106915773
Face book’s IP address.
2a03:2880:2130:cf05:face:b00c::1


I want one with b00b and 69 in it.
Also, should have used base36 instead of hex. Retards.

The only people in my experience that go against ipv6 are fresh-ccna-monkeys, who failed to learn it or 60 year old boomers, who are incapable of learning a new thing if their lives depended on it. We don't even have public ipv6, yet I managed to switch the entire company along with external equipment to ipv6

>>106915787
>IPv6 is unironically more complex than just using NAT
see >>106915646

>>106906949 (OP)
I'm sorry, it's just better to see "hey bro, jump on tf2 at 69.69.69.69" rather than "oh you need to connect to sijskojdofidsjosjdogidjsiogsoigjodijosdigjoiejgjeajgdgfsoijgdfoijaogjhae9rjgodfiajvodjogÑ:::SÑAÑ::::AÑ:AÑÑAÑAÑ"

Larger address space just means more spammers and bot IPs to ban

>>106915901
Nothing stops you from using
2a03:XXXX:XXXX:XXX:b00b:b00b:b00b:69
because the last four parts are for you freely to assign

>>106916191
>more spammers and bot IPs
That's what CGNAT is for.
Rather than banning an IP or and individual customer range, you can't ban the spammer at all, you can just ban everyone using that provider.

>>106915592
You can shorten them though
FC:DB8::1/64
FC:DB8::2/64
FC:DB8::3/64
etc.

You will never need a different subnet mask either unless you want ptp

>>106906949 (OP)
2001:0db8:85a3:0000:0000:8a2e:0370:7334
69.144.81.7

See the problem? If IPv4 was just extended with another octet or two for 6 in IPv6, it would look like this:

69.144.81.7.71.8

Yes it looks complicated but it's much better than this: 2001:0db8:85a3:0000:0000:8a2e:0370:7334

>>106916232
I can already do that now

>>106916265
You can not ban an individual behind a CGNAT on IPv4.

>>106916255
>2001:0db8:85a3:0000:0000:8a2e:0370:7334
even this imaginary IPv6 would be written as
2001:db8:85a3::8a2e:370:7334
and because per standard every IPv6 customer has a /64, it would be written as:
2001:db8:85a3::

>>106907869
No, but it sure works as one by blocking glowies and jeets from port-scanning and mapping your entire network.
That's why they hate NAT and keep trying to kill it by pushing the pozzed IPv6 on everyone.

>>106916255
>If IPv4 was just extended with another octet or two
Then you would still have the problem of the utterly stupid IPv6 package structure.
An IPv4 header is anywhere between 20 and 60 bytes with fluff that nobody needs.
An IPv6 header is 40 bytes.
And an IPv6 package allows a higher payload as well, so you have to send headers less often.
You can't even imagine just how much IPv6 simplifies things. It start with the very basics.

No modification of IPv4 can possibly make it good. No matter how much incompetent boomers seeth and cry.

>>106916302
>utterly stupid IPv6 package structure
IPv4*

>>106916289
>That's why they hate NAT
they LOVE CGNAT, anon.
They love NAT so soo much that they implement it on your ISP level as well, just because.
>port scanning
Nigger, IPv6 gives you a whole \64, so someone port scanning you would have to scan 1209239414263882237083648 ports.
Good luck.
Meanwhile every single IPv4 on the whole internet gets scanned by any random discord kiddy every hour.

>>106916269
Where did I claim that?
This was in direct response to banning everyone using the same provider.
Which you can already do in IPv4.

>>106916332
I was pointing out that you have to ban the whole provider and can't ban an individual.
On IPv6 you can do both.
You have less options on v4.

>>106916343
Which doesn't do you any good because spammers can quickly change to a million other IPs since there are so many. So you have to ban whole ranges anyway.

>>106916331
>a whole \64
A good ISP will give you a /48, which is 1.2 septillion addresses.

>>106916279
See it makes no sense, might as well give people a sha512 hash for their address.

>>106913570
jfc now i know why internet used to be good, racism is the only way

>>106916351
You ban the /64. Banning the individual IP makes no sense indeed.
Your banlist becomes an unsigned 64 bit integer. You don't care about half the IP.
And you can run dual-stack with no effort, because the 32 bit IPv4s can still be mapped within it.

Meanwhile what do you do with a spammer on CGNAT who shares the same IP as a hundred legit users? What do you do with a spammer who rerolls IPs with airplane mode on his phone?
The thing, that you think to become a problem with v6, is actually becoming much easier.

>>106916377
>it makes no sense
So 2001:db8:85a3:: is harder for you to remember than 69.144.81.7 ?
Consider that it gets even simpler, the first four digits are almost always the same,
2a03:: is RIPE, so all Europeans.
2001, 2607 and 2600 are ARIN
So you realistically see only those four. So in your random fantasy IPv6, you only have the db8:85a3 to remember.

Do all those simplifications make it confusing for you and you would rather really only have random digits?

>>106916415
>What do you do with a spammer who rerolls IPs with airplane mode on his phone?
How is IPv6 helping there?

>>106916289
>but it sure works as one
it doesn't.

>>106916519
You don't get a new \64 on reconnect.
IPv6 doesn't do any of those funky things and a ISP has no reason to do that.

>>106916533
>You don't get a new \64 on reconnect.
says who?
Why wouldn't the provider just grab a random free IP address for a new connection instead of bothering to keep IPs matching some MAC (which is easy for the spammer to change anyway) or something.

>>106916547
>says who?
The will of the ISP to make money.
It is cheaper not to do it. And there is no reason to.
>Why wouldn't the provider just grab a random free IP address for a new connection
Your connection is already linked to a customer account because the ISP monitors your data usage even when you have an unlimited plan.
Your modem, be it DSL or fiber, already gets a static IP and then builds up a VPN like connection (just PPPoE on DSL) which then gets it the public IP. This still happens even when you have a static public IP and no CGNAT. There is always some customer bound IP in the chain.
Those are the hoops your ISP has to jump through with IPv4, those things don't exist anymore with v6.

>>106916586
>Your modem, be it DSL or fiber, already gets a static IP and then builds up a VPN like connection (just PPPoE on DSL) which then gets it the public IP.
The fuck does my mobile provider know about what my router does with my landline provider?

>>106916606
>but what if its not fiber or dsl but a phone
They already got you mapped with your IMEI. Still cheaper not to rotate a pool.
You are only arguing about edge cases where in the worst case, it will just be as hard with v6 as with v4, while all other benefits still stand.

>>106916679
>Still cheaper not to rotate a pool.
How? Why?
Picking a random ass IP seems to be the simplest possible solution there is.

Maintaining a database with recently disconnected IP assignments and associated identifiers is extra work.

>>106916691
how do you make sure that this random ass IP you picked isn't already assigned to another customer?

idk bro, it seems like you only write for the sake of writing something

>>106916718
>how do you make sure that this random ass IP you picked isn't already assigned to another customer?
Because you do need to know the assignments of active IPs in any case.

But what you don't need to do, is to keep track of the former assignments of disconnected IPs. It is just extra work.

>>106916750
Ever heard of a hash list?
>what you don't need to do, is to keep track of the former assignments of disconnected IPs
You do, its a legal requirement. Which obviously takes up more space when you assign randomly.

>>106906949 (OP)
>>106907859
They could have just used x.x.x.x.x.x addresses but they had to absolutely overengineer it so every atom in the universe can have it's own IP address

>>106916900
No, it's not.
You just need to append to some write-only log every connect and disconnect and hand it over when police asks for it.

No need to try to match new connections with old assignments and decide on strategies on when to expire old assignments and so on.
It is extra shit you have to do. It's not the hardest thing in the world to do but it's still extra work that the ISP doesn't benefit at all from.

>>106913570
>>106913590
>>106914746
>>106916402
Yet the internet was ruined by AMERICAN companies and commercialization long before any jeet was even online

>>106907319
you'd be the first person I try it on
>>106907782
precisely zero. the only place ipv6 is useful is in direct p2p.

>>106917189
direct p2p sounds based and freedom pilled
fuck fagman, fuck stun, fuck turn, fuck kikes

>>106917210
I had my answer to it directly below but decided to keep it to myself. regardless, the status quo of ipv6 for p2p (and garbage sites that do captcha when on cell/cgnat) and ipv4 for web services is fine as it is

>>106917210
with zero anonymity

>>106916279
and 192.168.0.1 would be written as c0a8:1

>>106917235
wym no anonymity

>>106916302
only boomers want ipv6. otherwise it's some autistic troon from eastern europe or sonething.

>>106917267
cause its direct p2p with static unique IPs for everyone?

>>106917270
True, boomers want an internet connection, xoomers/yoomers/zoomers/alphoomers just want an instagram connection.

>>106917365
yeah yeah. at least you get it now, mister dinosaur

saar

>>106917478
India super power

>>106907859
>>106906949 (OP)
im not going to remember 90-34290-34.23.42.34.23.5235.4234.23.4 fuck that. fuck off.

>>106915646
>NOOOOO IT'S NOT IPv6 THAT'S THE PROBLEM IT'S SOME MADE UP SHIT THAT I IMAGINED HAPPENED WHILE ON CRACK
Why are IPv6 apologists always like this. And yeah I stopped reading your imaginary story in the middle of the second sentence. Nice blog post nigger.

>>106913076
>Switching to another ISP for shit like this should be the normal behaviour.
>no nat, restore end-to-end connectivity
>a simpler configuration process.
and then everyone stood up and clapped
literally communist tier delusions

>>106917589
>>106917578
it's ok, boomer, you can keep using your DOS box,
now take your meds and stop rambling with no reason or argument

IPv6 is GAY and RETARDED

>>106917589
3 years ago I had my old ISP close out its business because of a catastrophic failure which they didn't disclose.
I switched to another one but saw that i didn't had IPv6, I kindly asked for a subnet and that I was willing to pay extra but they couldn't even give it to me, one of their technicians said all of their network is IPv4 only. A month later I switched to my current ISP that has top tier IPv6 support with a static /56 subnet for regular customers and one or more static /48 subnets for businesses.

>>106907319
for something to be illegal
it must be possible
give it your best try

>All the retards saying IPv6 bad don't even understand how it actually works
The absolute state of /g/

>>106915015
if your company disables privacy extensions, they also require you to use a vpn and likely use a tls-inspecting firewall
ipv4 wont help you there
>>106917276
>static
you have to go out of your way to get a static v6
you have as much control over your v6 prefix as you do a v4, you are entirely at your ISP's mercy
when i had google fiber i effectively had static v4 and v6 because they never changed, the municipal isp i currently have rotates both weekly
>>106908784
its like the credit card industry in the US
theres a reason insane CC rewards only exist in the US: no other civilized country allows payment processors to rentseek 2-5% (or 10% if youre in adult industries) of every tx
but it will never change because consumers and banks have gotten fat off the "rewards" programs at the expense of small stores that pay higher fees due to being unable to negotiate large discounts
>>106912501
my company switched to v6-only internally with dual-stack externally and it solved so many problems
we exhaused the RFC1918 ranges in the 00s and eventually had to setup multiple NATs within our internal network
it was nightmarish to maintain
so sounds like a skill issue on your end

We need to switch to alphanumeric representation for ipv6. That's 10 alphanumeric 6-bit values [a-zA-Z0-9!*] and 1 hex 4-bit value for each 64-bit half. Then we can name our servers
> GayWebZone7:CP4Fedz!!!0
I don't see a problem with this and everyone can start right now

>>106918601
This will finally end DNS

>>106917504
it's closer to externalIP:<2zeroIPs>:localIP
not that hard to remember

>>106908702
i too hate the ssl and its anal retentive ecommerce security thats been jammed into everything

>>106917235
Because hiding behind your router meant something appearantly.

>>106916946
Well that was an interesting dialog.

>>106916168
Tell yourself that when you get a /65 instead of a /64 and want to support Android (which doesn't support DHCP for IPv6).

>>106919196
Also imagine if you got a /64 but wanted to have separate subnets for whatever purpose. lol not supported, you NEED to have a full 64 bits free for SLAAC to work, the only way for that to work is to use NAT64 and rely on private IPv6 addresses, in which case you get the annoyance of NAT and the piece of absolute shit that is IPv6, and at that point just disable it and use IPv4 only.

>>106919227
>NAT64
Actually this is different than just NAT, sorry my head got garbled for a second

>>106907859
>an IPv6 usually lasts 1 to 5 hours if not statically assigned.

False. Last time I checked it didn't change for weeks, it only changed when I restarted the PC.
IPv6 is malware.
Never use IPv6 if you care about privacy.

You can't do vlans with IPv6 so it's useless

>>106916331
You can scan the entire IPv4 range in 5 minutes with the right config https://zmap.io/

>>106919830
weak bait

>>106915436
Amazing how the internet is getting worse with time.

guys why didnt they use alphabetical notation instead of hexadecimal for ipv6? if it was alphabetical you could create addresses that were memorable and legible.

>>106914363
>>106915436
sounds like CGNAT.

How can I tell what my isp gives me like a /56 or /64

>>106918711
We have a few legacy machines that require us to go in and manually add SSL certs to them, so we have a couple of self signed internal certs set to expire in the 2060s. In delusional spec lala land this is terrible practice and you shouldn't do that. You should implement some automated deployment to these machines and redeploy your certs every week.

OK... I get that that's what the spec people want, and what we're doing is retarded, but the protocol these machines uses is hacky and you can't just SSH or telnet in. The other major concern is that the eMMCs that they use are /insanely/ fucking bad. If we switch to pushing new certs every month, the flash is going to die and then someone (probably me) is going to have to pull the board out and replace the eMMC. You literally can't get an up to date version of the firmware stored on these pieces of shit, so we'd have to rollback 3 or 4 chips on the boards to the last known full version release just to replace the eMMC holding the certs, then go through and do incremental upgrades to the current version, which has a decent chance of cooking other flash chips on them. Realistically you'd want to pull and replace all of the chips, and probably a bunch of questionably ancient capacitors too while you're at it. Technically doing any of this process would be illegal, but the company that made these pieces of shit hasn't existed for nearly a decade and there's nobody left to care.

Theoretically you could pull a working board now and dump the contents of all these chips, but there ain't a snowballs chance in hell you're going to convince anyone that that's a good idea. The current plan is to just do retrofits the next time something breaks because you literally can't get replacement parts anyways. We're not going to do that until something breaks though because we don't explicitly need all 3 of the machines up at once, and the retrofits will run close to 80k per machine.

And yeah, none of these are ipv6 capable.

>>106919248
That's the thing it isn't even NAT64 or NAT46 it's NAT66 which as others have already pointed out is literally a worse fate than I have no mouth and I must scream.

>>106922185
>I get that that's what the spec people want, and what we're doing is retarded, but the protocol these machines uses is hacky and you can't just SSH or telnet in.
I mean if it worked I would want to live in a textbook communist society too, problem is we live in reality and not some Hollywood musical.
It's funny but these discussions always immediately show who's ever worked with this shit before and who's just larping because the ratio of devices capable of modern automated certificate renewal are laughably small.
But don't worry these retards will push for 60 second SSL certificates soon just ignore the fact that stolen SSL certificates are completely irrelevant and the only certificates that ever get abused are the ones used to sign drivers, thank god those come with 30 days forced expiry right?
This shit is just as delusional and disconnected from reality as people insisting companies change their 8-9 digit Dollar value production machines, investing another 8-9 digits on a new machine + the cost of downtime and lost customers + renovation costs (because these things are NEVER drop in replacements) just because hurr durr it relies on W95 or a serial port or because it can't handle the newest faggoty ass change to how many femtoseconds a SSL certificate should be valid for.

>>106906949 (OP)
Eternal reminder that the """"people"""" that designed the IPv6 spec designed it with the assumption that the IPv6 prefix for literally every person on earth would be statically bound to that specific person forever in perpetuity even if they change ISP or even country or continent but didn't make ANYTHING of that sort mandatory. All they did was shit out a RIPE recommendation no one gives a fuck about that came literal decades too late to possibly make a difference.
https://www.ripe.net/publications/docs/ripe-690/

Why would private citizen prefer to have the same internet addresses for an extended period of time? It never made sense to me as it only brings legal disadvantages

>>106922813
nice try fagman

>>106914753
I'm only a CCNA right now, but I'll be getting that CCNP soon. Lets keep this industry a secret so these stupid programmers starve.

>>106922610
>This shit is just as delusional and disconnected from reality as people insisting companies change their 8-9 digit Dollar value production machines, investing another 8-9 digits on a new machine + the cost of downtime and lost customers + renovation costs (because these things are NEVER drop in replacements)

Don't forget the part where if you have to crack open a wall to move a piece of equipment in or out nothing is ever the same afterwards. That machine near the hole in the wall that was installed in 1993 will magically stop working perfectly right after it got fresh air for 3 days, even though it was a bone dry middle of July day. Every single test you do says that the machine is operating perfectly normally, but for some reason your reject rate on half your jobs went from .1% to .15, and it's costing you 17,000 a month in waste. Gremlins and ghosts like that are hardly the main reason, let alone the only one, why companies hate change, but it's a solid enough percentage to be worth mentioning.

Things would be so much better if developers and spec writers had to chose between letting a 300CU argon tank roll over their foot without shoes on or use their bullshit on a 10 year old base model laptop for a week. Cool, your a dev working for a company that buys new workstation laptops ever 2 years like clockwork, and now you cannot even conceive of having to worry about pre N wifi devices shitting up your network. Meanwhile in the real world I still have 3 wifi B and G devices that I'm finally getting rid of sometime in December (and realistically it's going to be February). Do you have any idea how fucking happy I was to get it in ink that I'd be able to turn that shit off and crank the minimum broadcast rate up to a sane levels? Literally 70% of our wifi bandwidth in parts of the building is beacons being sent out at dialup speeds because most old hardware will "play it safe" and use the minimum allowable speed for that.

>>106906949 (OP)
>tfw IPv6 is the entire reason I'm able to homelab at all with my ISP because IPv4 is so locked down on their hardware there isn't even port forwarding in the modem interface.

I am able to use IPv6 with tailscale to access every machine on my network anywhere in the world as if I were home on my LAN, and it's so fast that even over a cell connection in the middle of bumfuck nowhere, loading content feels as fast as loading it locally on the server itself. I don't know what kind of fucking wizardry they're using but I'm having a better experience with 500Mbps/15Mbps 5G home Internet than I did with 1Gbps/1Gbps fiber with wireguard to give me access to my entire LAN.

>>106924752
It could be that your ipv6 address is routed differently.

First time trying to use the steam marketplace this week. I sure love sharing ip with thousands of melanin rich people.

I've eaten bans meant for other people a lot.
Being on Dual Stack-lite without a real IPv4 address is pain. My IPv4 address is the same as the other users of my ISP.
Why doesn't 4chan accept IPv6 connections?

>>106912043
set up a radius server
put it on an edge blade make your lan a vxlan with sd-access for fun ipv6 only just works
get a wireless lan controller with ansible scripts to restart web servers next door
ask your neighbors if you can dig cables to their house make cdp neighbors with them locally
make a nat pool overload one public wan address for the neighbors
get a clump of cable the size of a chubby teenagers thighs in a port channel routing to the ISPR cone
get hoa funded ap mesh antennas on the rooftops
do not allow spectrum to visit you

>>106906949 (OP)
ipv6 still gay +ugly
not my problem

>>106908702
Based IT realist

>>106912007
its not just boomers
its everyone that isn't IT
including Database Administrator IV who doesn't understand why adobe reader can't be used to edit pdfs because an update changed her default app. these execs like servicenow ceo saying AI is going to replace their workforce is extreme wishful thinking+optics, middle managers do not want anyone fucking with their department period.

>>106916251
You wrote this like that was easier, it's not. I can't remember what the address is since looking at it before I started typing this.

>>106926036
>>106912007
>boomers dont adapt easily
a buddy of mine runs runs a legit boomer family small business He asked me to help with an email issue, I discovered that in the year 2025 he was physically printing invoices and scanning them as jpegs to email to customers instead of printing to pdf. He literally did not know that was possible. AI is never going to help him with his business and even if it could he will NEVER learn how to actually use it to his advantage. The degree with which software companies are out of touch with the majority of their users is insane.

>>106926180
FC00:0DB8:0000:0000:0000:0000:0000:0000/64

>>106924752
why doesn't your isp let you configure their modem in bridged mode?

>>106924667
>Literally 70% of our wifi bandwidth in parts of the building is beacons being sent out at dialup speeds because most old hardware will "play it safe" and use the minimum allowable speed for that.
God I fucking hate mixed networks. Wifi is always a pain in the ass anyways but fuck me the selection of what gen any given device will use is so fucking god awful.
>That machine near the hole in the wall that was installed in 1993 will magically stop working perfectly right after it got fresh air for 3 days
Yeah I've seen something like that happen before too. I worked in a chemical lab ages ago, one day a pipe in the ceiling just gives after a few decades of service and we have to move a bunch of equipment, none of them getting any dust, debris or liquids on them mind you, our containment measures worked as intended but we were paranoid.
We move a half million Euro brand new GCMS machine literally two tables to the right and it stops working. Diagnostics run fine, there's nothing obviously wrong with it. When we put it back in its "old" spot it just started working again.

>>106924752
Packet routing is way more efficient and you make less hops to reach your desired destination. That is the reason why.

>>106925310
Moot actually added support for IPv6 10 years ago, but he was a midwit and implemented a shitty system that banned individual IPs rather than IPv6 subnets, the same as IPv4, so instead of fixing it he scrapped the whole thing and gave up on it.

>>106925334
What?

>>106906949 (OP)
they fucked up and made ipv6 harder than all that bullshit. just not intuitive. i'm waiting for ipv7

>>106927980
>God I fucking hate mixed networks. Wifi is always a pain in the ass anyways but fuck me the selection of what gen any given device will use is so fucking god awful.
Tell me about it. The worst part is that old devices don't support roaming properly, if they support it at all. In that one section of the building I have 3 high density APs and I'm fully using the 2.4 GHz space. The entire facility has the 2.4Ghz channels based around that one fucking corner because we have so much shit there that can't roam and needs to have pretty stable wifi over a decently sized area. Everything except 5 or 6 devices supports beacon speeds in excess of 12Mb, but we have dozens and dozens of devices that will always run at the minimum available one.

It's going to be years before we fully get rid of 2.4 only devices, but I'll be able to make things so much fucking faster and more stable if I can set the beacon rates to 6 (or hopefully 12) and the frequency to 200+ms instead of having to keep it at 1Mb and 50ms polling intervals.

I've already done this in the office parts of the building without telling anyone. In retrospect maybe I shouldn't have done that so the pain would have been more obvious, but I just couldn't live with doing that to the engineering team.

>>106928048
Capital One got an IPv6 /16 or something stupid like that not too long ago, so it sounds like we're back to the IPv4 routing clusterfuck situation already. It sounds like route aggregation was a pipe dream all along.
I'm not a network engineer so I don't have to worry about any of this nonsense in a professional context but it sounds really dumb.

>>106912501
Based and correct. Only troons and hobbyfags who never configured an enterprise firewall in their life shill IPv6. No, your pfSense shitbox does not count, unemployed nigger.

>>106908702
TRVTH NVKE

>>106913582
>"You want your own private IPv4? 99$ a month plus tip."
That's how it should have been from the beginning. This would save countless IPv4 ranges and the only "people" complaining about this would be unemployed hobbyist lintroons who absolutely NEED to wireguard to their Plex (running on a raspberry in their closet) when at some lintroon convention, to show to other fellow lintroons that THEY ARE 1337 H4XX0R5 TOO! The opposite of what is /g/ consensus on IT should always be implemented.

>>106906949 (OP)
I'm a white man, I disabled IPv6 in my router

>>106907015
Fpbp
IPv4+NAT accidentally created the most sensible model for thinking about networks in a secure way. Having private address space behind a router with port forwarding simply forces you to think about networking in a way that allows you to clearly understand what's visible to the internet and what isn't, with minimal mental overhead, whereas ipv6 makes it overcomplicated for no reason