IPv6 isn't as big as you think.You need a /64 for each device so it can subnet within virtual networks, then each router usually demands a /48 per household. Combine that with next generation IoT devices and the fact that only IP addresses beginning with 2xxx: are allocated then you soon realize that IPv6 tops out at around 17 trillion devices, which isn't much when you consider all the microcontrollers that want their own IP.Until variable length IPs become the standard we will keep having to rewrite the IP stack.
yes dad
>>107215411>You need a /64 for each device so it can subnet within virtual networksIPv6 noob here. Why was this a thing? I dont see the usecase for each device requiring that many addresses and throwing away half the available address space doesnt sound like a good idea and we end up with insufficient /48 subnets and go back to using NATs
what should my ip r output so i'm sure i'm utilizing ipv6? or does it not matter?
>>107215411Ignoring all the other inaccuracies in your post, I don't think you understand how big 17 trillion figure is.
>>107215411By 2030 there will be 17 trillion indians all clamoring for IPv6 addresses.
>>107215736Nanomachines, son!
>>107215875haha, I look at xkcd too!
>>107215411IPv6 was a mistake.>ff06:0:0:0:0:0:0:c3this is not an IP addressNAT is a hack solving IPv4 pretty well meanwhile saving normies not having firewalls.
>>107215411My ISP only partially supports IPv6. Works totally fine on their lower speed plans, but is 100% disabled/broken on their fastest speed tier. Not sure if I should really give a shit or not, at the moment it doesn't seem to matter much, but in the future more shit might be IPv6 only i guess and that could cause some problems for me.
ipv4 is enough for my house, and i have a lot more than average connected devices in this house.
>>107215411They should have kept NAT.
>>107215411Exactly. This is what I've been saying for years.They fucked up by restricting themselves to only 128 bits.If they had actually committed to a useful number like 512bits (could even get away with 256) it would have been adopted over a decade ago. As it stands, its barely an improvement over IPv4. Esepcially when considering the fact that all existing hardware and network software configurations need to be adapted to work with IPv6 for questionable level of benefits.
>>107215551>half a the available adress spacewe are throwing away 99.99999% of the available adress space. Each bit is a doubling of the space
>>107215901>pay more money>lose IPv6 accessuhhh based?
Okay hear me out. We need a simple script that converts an IP address of type 255.255.255.255.255(of whatever length is necessary) into an IPv6 address. Then when we have this, it's always used and everyone forgets IPv6 looks the way it does.
>>107215411The problem with ipv6 was they didn’t think about what the transition would look like. Protocols as fundamental as ip never die, they accrete like sediment. They needed to design an ipv6-over-ipv4 protocol for the transition period. They also needed to design a way to give fake ipv4 ips to client software to force them to be compatible with ipv6. If they had, ipv6 would have won in ten years.
>>107216944There are IPv4 mapped IPv6 addresses for this exact reason. You cannot map an IPv6 address to an IPv4 address without losing a significant portion of the address and then you end up in heirarchical routing weirdo land at best or worst it simply breaks. Most of the transition mechanisms are catagorized into mapping and tunneling mechanisms. MAP-E, XLAT, 6-to-4 are mapping and MAP-T with some VPN based mechanisms like 6PE and 6VPE or IPv6 tunnel brokering are tunnel based mechanisms. IPv6 fixes most if not all the retarded issues of IPv4 literally the only problems with it come from SLAAC and DHCPv6 that isn't strictly slaved to only the DHCP server. If you are filtered by hex sell your computer consoomer tier thirdie.>>107215969IPv6 has NAT.>>107215411Giving a /64 to a home router works just fine because your home router doesn't work like a normal router where on either side of it sit two distinct networks with two prefixes. Look up Broadcom distributed switch architecture. Devices just get or generate an address from the /64.
>>107217052IPv4 mapped lets ipv6 clients support ipv4. I’m talking about the other direction, how do ipv4 clients support IPv6? It’s an absolute necessity. > You cannot map an IPv6 address to an IPv4 addressYou create a vpn, and map IPv6 addresses to IPv4 ones within that network. You give a separate network to a given process, so it can’t share addresses with other processes, it needs to share domain names. Some programs will break but 99% will work fine.
>>107215878I don't get the xkcd hatred, munroe is clearly much smarter than me much less mouthbreathing dysgenics from 4chan
>>107215875Ah yes just like my cock they harden in response to physical trauma.
>>107217174they support it by not IPv6 and the network does the rest. Unfortunately both network people and devshitters have both agreed the network needs to handle most of the complexities and hosts should understand as little as possible. All designs for transition are based on IPv4 hosts being blind to IPv6 and as far as they know IPv4 goes into the network and out comes IPv4 even if the whole thing is IPv6. This is what 4-6-4 does in DNS and in combination with any of the dataplane mapping or tunneling techniques IPv4 only hosts can access an IPv6 only network.>and map IPv6 addresses to IPv4 ones within that networkthis part is just 6-to-4 NAT which is fine. I was talking directly encoding an IPv6 address inside an IPv4 one which will not work. You can encode the other way and it works just fine.https://www.rfc-editor.org/rfc/rfc4291.html section 2.2also relevant https://www.rfc-editor.org/rfc/rfc6052 section 2.
>>107215411> variable length IPsWe already had something better.Like IPv6, variable length IPs are not needed.We will switch back to uucp bang paths to navigate multi-level NAT.Single-level “URLs” and “e-mail addresses” were always scams to try and make individual devices spy-able on the open internet by directly addressing it to contact the logic bombs, bugs, back-doors, etc. embedded in the device.
>>107217052>IPv6 has NAT.It does? When did they add that?
i request a /60 from my isp and it's enough for me
>>107217394They clearly put a lot of thought into it. I hate to say it, but the only thing I can think of that they did wrong is making the addresses too damn long.
>>107215969why do you want nat
>>107217208wasn't hate
>>107217497It's an answer to OPs concern that we are running out of ipv6 addresses. I liked it because I found it easier to create and manage networks.
>>107216874There are lots of places in IPv4 where they could have used extra, optional byte (or many bytes) in the IP header to expand it and make it 256 times bigger. Easily and trivially.Kind of like phone numbers, but with an extension.A lot of user-level identifiers (like URLs) support adding a port.But they didn’t, and they wasted much of the available IPv6 space and basically implemented all this insanely and intentionally complicated bullshit.
>>107217492> They clearly put a lot of thought into itUnfortunately, they were literal retards, or malfeasors (probably both). The world would have been better-off without them, and IPv6, existed.
>>107217500put more effort into thsi
>>107217440> it has that??It probably has everything you could conceive of.That’s the problem.I hear USB 5.2 will begin support serial data again, instead of being a bulk DRMed power delivery system. I haven’t plugged anything into my phone in years, lol.
No excuse to not at least have dual stack on your network
>>107215411I’ve never had or met anyone that had an IPV6 address personally.
>>107217743I have an entire /48 'cause my ISP is good. My country's regulators are also pretty based requiring P2P instead of P2MP single mode fiber installations, which massively increases bandwidth in the future too.
>>107218283>My country's regulators are also pretty based requiring P2P instead of P2MP single mode fiber installationsWhat does that mean? One dedicated fiber/transceiver per customer instead of this gay ass shared medium PON bullshit with passive optical splitters in between? I believe ISPs love that to bits because shared medium means free traffic shaping they don't have to call traffic shaping.
nat is a cisco invention. total nat death. the internet is supposed to be connectable.
my ISP provide IPv6 as dual stack (ipv4+ipv6) via DHCPv6+SLAAC (IPv6/128 address on WAN interface and IPv6 PD (IPv6 Prefix Delegation)/56as I understand, WAN IPv6 address will be static?What's the downsides of this implementation? any possible attacks?
>>107220554That is the proper way of doing itThere are no downsides
>>107220554>What's the downsides of this implementation?you can no longer reset your router to avoid IP ban
>>10721555164 bits is half of a 65-bit address space. It is 1/264 of a 128-bit address space. There are 4 billion 64-bit subnets of IPv6 for every single IPv4 address. That's why we'll never run out. Human population will be max ~10 billion, enough for every person alive to have 3.5*1028 addresses.
>>107215551>>107221900Fug my numbers broke. Was meant to be 2^64 and 10^28.
>>107215411Take every atom composing the earth Earth, divide them up into groups the size of a grain of sand. You can assign every grain of sand 122 million IPv6 addresses. How do you think we'll run out?
>>107215411ITT networklets again:>make wild assumptions about ipv6>claim that nat is secure or private>call ipv4 more readable than ipv6>defend greedy ISPs selling ipv4 addresses at 4000% markup
>>107215411>You need a /64 for each deviceSLAAC was one of the most retarded ideas ever conceived. they went to the trouble of expanding the address space and then wasted it>>107220554wasting lots of address space, the very thing that IPv6 was supposed to obviate. They said "oh, 64 bits is plenty, we can afford to waste lots of them!" just like they said with the 32 bits of IPv4. And all of a sudden when you recommend that ISPs hand out /56es (if not /48s) to anyone who asks, they start to think address space is scarce, the WHOLE THING YOU WERE TRYING TO PREVENT. My ISP won't do a prefix delegation at all, you get one (1) /64 and that's it. Guess why? There should not have ever been such a retardedly huge minimum subnet size. If they gave me a /96 and let me subnet that, I'd have plenty of addresses. And so would everyone else. But nooooo, fucking SLAAC and it's pointless bullshit /64 minimum subnet size.
>>107215411Ip6 subnet is a lie, it's just a convinient way for expressing routes in situations more complex than "send all shit via default"
IPv4 is kept around to charge outrageous prices for IP space. True freedom is ipv6 where space can last a life time.
>>107215411What is IP√6?
>>107222279IP2.45
The only reason why IPV4 "ran out of space" is because massive chunks of addresses were given to useless American companies and government agencies. Combine that with India wanting more IP addresses to scam people with, and you run out.You only need enough IP addresses so that each household gets 1 IP address. Having 1 IP address per device is a retarded idea and it's bad for security. Imagine giving direct access to all your security cameras, IoT fridges, network-enabled salt shakers and subscription-service-based mattresses.
>>107222399To what household does your phone's mobile data belong?
>>107222412All phones on a cell provider can just share a single IP address because they don't belong to a "household" anymore.
I disable ipv6 everywhere. Anyone who wants to make fun of that: I dare you to disable ipv4 for 24h and report back to us.
>>107222101Ok, i get it. Fuck v6 scam shit.Thanks
>>107222439You are retardedPlease never post on a technology board ever again
i just can't get over how ugly IPv6 addresses are. why did they have to use colons, what's wrong with using periods?
>>107224481ok, woman
>>107219543yes, no splittersthey obviously reduce bandwidth per connection as seen from your end (at home) divided by how many they're merged intoand apart from that the splitters (from your end they look like mergers I suppose) aren't even that cheap meaning they're only worth it if overused to shit, which also will stop more than 1-2 ISPs from being profitable in your area. The second or third will already have too few customers to make the splitter worth it if fees aren't crazy. p2p? ISPs can offer good service at pretty competitive rates even with a small number of customers.laying fiber p2p is the way to go
>>107225556Far too expensive to deploy at larger scale.It's great if your country can afford to subsidize that, but I wouldn't expect most countries to bother.P2MP is just far more economically viable for residential internet.
>>107225625Works absolutely fine country-wide and there really isn't any scaling beyond that, you can do it in all countries.Inb4 fiber wires are too expensive. They're not and they're easier to install than electrical wiring too, you can have one cable with hundreds or even thousands of fibers inside and no one says you have to do only one cable etiher.Anyone who does it another way and it's not a temp aboveground installation is fucking up.
>>107225625NTA but I think the other anon is Swiss, just like me. P2P isn't "way more" expensive. It's a bit more expensive, that's true, but not by a big amount.>P2MP is just far more economically viable for residential internet.Maybe. But then you as the customer can't choose freely between ISPs because they all give you the same speed (meaning that the competition is nonexistent). With P2P, an ISP can have speeds that another ISP doesn't provide. E.g. Init7 from Switzerland does this: they provide 25 Gbps for 70 bucks a month. And mind you, Switzerland usually isn't cheap. Try to get that price in another country.Pic related, it's from an Init7 PoP.
>>107217687my isp does not support ipv6
>>107224481They should have just been base36 or base62 strings instead of the fucking mess they created.
>>107226892Lol, sorry but you're just wrong.It isn't a "bit more" expensive.> In most typical scenarios, the total Capital Expenditure (CapEx) for P2MP GPON is estimated to be 30% to 50% lower than P2P EPON (Active Ethernet) for large-scale residential deployments in medium-to-high density areas.
>>107227110No problem :) Hurricane Electric gives out /48s for free: https://tunnelbroker.net/
>>107227423>THINK OF THE SHAREHOLDERSReminds me of the old Onion sketch.>Week in and week out you come back like a house broken Pomeranian, loyal to the bitter end. Good boy.
>>107226892>>107225692You're ignoring how much your government is paying for that fiber.~60-80% of the cost for that fiber was paid by the government, not the ISP.So sure, if you ignore how much is ACTUALLY being spent on your network, you can pretend P2P is magically cheaper to deploy because Init7 doesn't have to pay for it.
>>107227423>>107227479>this is the best cope mutts can come up with when confronted by 25G symmetrical FTTH for less money than their 50/10 coax
>>107227501Feel free to continue pretending to be retarded to make your argument look better, but all you really do is make yourself look retarded.Even the Swiss government admits P2MP is cheaper, they freely admit it and estimated that P2P cost them an additional ~$500M.https://blog.init7.net/en/hochbreitbandstrategie-glasfaser-ausbau-soll-finanziell-unterstuetzt-werden/>Compared to the P2MP 1-fiber model, an additional CHF 400 million is required for the P2P model – but this model is the only way to ensure fair competition between the providers. And again, i'm not arguing that the Swiss method for doing fiber is BAD, what i'm saying is you're totally misrepresenting the cost and hiding it behind goverment spending and pretending what the ISP pays is directly comparable to a private fiber network in the US being run by Google/AT&T/Verizon/etc.Except the networks in the US are paying out of pocket for the network infrastructure and the Swiss ISPs aren't paying that at all.
>>107227479>You're ignoring how much your government is paying for that fiber.nta but I don't even know what point you're trying to makehow is it relevant how much a FTTH infrastructure is subsidized?the only reason why it would even make sense to bring it up is if ISPs were still charging a shitload of money for mediocre throughput, which clearly isn't the case here
>>107215893NAT is implemented by a firewall, retard. You can do NAT6 in IPv6 too if you really want to, in fact it's sometimes the only solution to problems like wanting to assign IPv6 to your Wireguard tunnel without assigning statically allocated global addresses.
>>107227630He's arguing P2P "It's a bit more expensive, that's true, but not by a big amount", when it isn't "just a bit" more expensive, it's QUITE a bit more expensive. P2P is demonstrably ~30-50% more expensive, he's further going on to say that Swiss ISPs are able to offer faster AND cheaper internet than in the US when the ONLY reason Swiss ISPs can offer cheaper internet is because that P2P infrastructure is being subsidized. It ignores the true cost of the infrastructure and he's just handwaving away that BILLIONS of dollars invested as not a "real" cost because he doesn't see it on his monthly bill.Just be open about it, unless your government is ALSO willing to spend billions of your tax dollars to connect less than 10 million people to high speed fiber, it's not really something that can be replicated. Feel free to advocate for governments doing this, but don't pretend like that $4-5B cost doesn't exist at all, or that it isn't coming from your tax dollars. As my original point stands, >>107225625, P2P fiber is GENERALLY far too expensive to deploy at larger scale unless it's being subsidized heavily, P2MP is the only real economical way to deliver high speed fiber to residential customers WITHOUT relying on subsidies.
>>107217526You can install do NAT6 on your router and then just distribute a ULA down to your LAN.
>>107227685>the ONLY reason Swiss ISPs can offer cheaper internet is because that P2P infrastructure is being subsidized.Isn't that because of competition? Isn't US completely fucked precisely because telcos are allowed to establish cartels?
>>107224481They in fact went with both because IPv4-mapped v6 addresses are a thing.This is a valid IPv6 address:::127.0.0.1
::127.0.0.1
>>107227708>Isn't US completely fucked precisely because telcos are allowed to establish cartels?Yes. Also don't look up just how much said cartels receive in subsidies directly from the US government or it will make his seething about muh subsidies look even more pathetic.Especially don't look up internet prices or speeds in Hong Kong, seeing literal 10000x speeds available compared to what they have might just make them shoot up a school.
>>107227746~$42.5B for a country of ~340M peopleAround $125 per personvs the Swiss ~$5B for a country of ~9M people. That's almost $600 per person.
>>107227708Again, I am NOT arguing in favor of the US system, i'm asking him to objectively argue in favor of the swiss system WITHOUT LYING.
>>107227828where did you get that 5B from?
>>107217687>>107227431She's so cute bros
>>107227906Swisscom says they're spending around 1.5-1.7B CHF per year on CapEx the majority of which is fiber expenditures and they hope to have ~80% of the country covered by 2030.That alone will be ~$5B+.There is also the WIK Institute Recommendation that calls for 3B CHF invested for expanding fiber to cover remote regions of switzerland.And then another ~750M CHF from a federal source for a "kick-start funding" program for subsidizing the sparely populated areas rollout. Switzerland is pretty public about all of this shit. In any case the final subsidies will likely amount to far more than $5B, that's just an easy round number to throw out that I can fairly easily source and already demonstrates that Switzerland is subsidizing their fiber to a massive extent compared to the US.
>>107227975Oh over five years. Suuure. And I think you're reading it wrong, isn't the 1.4CHF the subsidy?
>>107228010Swisscom is publicly owned, anything they invest into the fiber network is a subsidy for the ISPs that will then use the network.
>>107228033>>107228010Also, the $42.5B figure for the US was over like 20 -30 years and covers shit besides fiber too.
>>107215439he sounds like a boomer, yeah.
>>107228047So, can you get the good numbers then?
>>107215411They lacked oversight when they created IPv6.The subscriber ID is held on only 16 bits which is very insufficient.They are probably cooking a 256 bits ipvX which extends IPv6 and acts as metadata, with a prefix of xx bits to identify if it is a person, organization, government, iot device etc, xxbits for identifying a type of relationship (owner, ISP customer etc.) and will contain xxx bits of personal identifier, immutable to a subscriber if it is a person, and of course, tied to digital ID.
>>107228033Swisscom been privatized for a while, they may get some money but not to the extend you assume. >tfw i get fibre this century
>>107228175*has
>>107228142I don't need to, the US has spent far less than $125 per citizen, Switzerland will likely end up spending far more than $500 per citizen.
>government spending money on infrastructure is le bad we must send further trillions to israelwhy are mutts like this? is it really just the lead water pipes?
>>107228538...it's almost like you're ignoring my posts entirely and just arguing in bad faith, i for one am shocked>>107227846>I am NOT arguing in favor of the US system, i'm asking him to objectively argue in favor of the swiss system WITHOUT LYING.You're arguing in bad faith by acting like the subsidies aren't real money being spent on this.The ACTUAL real world costs are still there and using government spending to cover the capital expenditure is fantastic for the consumers, but you can't hand-wave that expense away as if it doesn't exist JUST because the consumers don't see it on their monthly internet bill.It's still being paid for by taxes and other public funds. You are comparing US ISPs that are paying for the infrastructure vs Swiss ISPs that aren't paying for the majority of the infrastructure and then acting like the price you pay in comparison to the US consumer price is at all a fair comparison is the problem I have. Feel free to argue the Swiss system is better, feel free to argue P2P is better than P2MP. But don't argue that Swiss internet is CHEAPER, because it isn't; the costs are just paid upfront from your taxes instead of in your monthly bill and further don't argue that P2P is cheaper or marginally more expensive, when it's quite a bit more expensive to deploy than P2MP, especially at a larger scale. The Swiss government has just decided to bear the cost burden for the good of the Swiss public, which again, is fucking fantastic. I have ZERO problem with this.Just stop trying to pretend like the subsidies don't exist or can be fully ignored because its not listed on the bill you pay to Init7.
>>107227431thats not real ipv6. its just ipv4 with extra steps.
>>107222101NTA but can anyone refute this? I had similar suspicions but wasnt sure if its a lack of understanding on my part
Fine, I disabled ipv6 on my pc (not on my router) what now?
>>107231753why did you disable it? just turn it back on
>>107215411>You need a /64 for each device
> all the microcontrollers that want their own IP.Not with my money glowies.
>>107231786VPN leakage and /g/ propaganda
>>107222101>>107231712why the fuck do you need more than /64
>>107231823I need *less* than a /64, anon. I'm pissed that they go to the trouble of enlarging the address space with an incompatible protocol change and then waste way more of it than they should have, defeating their stated objective which is "there's so much space that we don't have to think about it". There would be if it weren't for fucking SLAAC, which was a stupid idea to begin with.
>>107215411True
Yeah I need a /23 but they only gave me a /43, so I told them if they dont atleast give me a /32 they will need to assign a /64 to their teeth to because they are about to become distributed devices
>>107215411>Until variable length IPs become the standard we will keep having to rewrite the IP stack.Variable address length is resource intensive. You have to check for length twice, once for the source address and once for the destination address, plus subnetting becomes basically impossible and/or an attack vector.Packets are restricted to 1500 MTU in size, there must be a maximum size for addresses length or adresses can actually be longer than the packet's payload size. Going from 32-bit (IPv4) to 128-bit address length (IPv6) is more than enough, 4 times the size means about 79 octillion times the number of addresses, which are about 340 undecillion in total.
>>107233415I'll agree that going for even more is stupid if things are allocated in vaguely sane manners, but some of your arguments are nonsense.>Variable address length is resource intensive.So fucking what? We have asinine compute power available for milliwatts of power thanks to how easy it is to shove random asics into everything these days. There's a bunch of ways to encode dynamic sizes into header information and shove that through hardware designed for it, with optional software fallbacks for things outside the range of what the hardware supports (or just because some chinesium switch wants to be giga cheap).>plus subnetting becomes basically impossible and/or an attack vector.How? This would all be sorted out at the switch/router anyways.>Packets are restricted to 1500 MTU in sizeActually it's 64k. Just that most devices and protocols can't support more than 1500. There's zero reason why the internet backend couldn't be forced to upgrade to larger packet sizes. It'd actually help with a lot of stuff as bandwidth rates continue to increase. Companies wanting to have large amounts of outgoing data could get equipment that could handle the larger packet sizes, and the equipment near the clients could fragment those into smaller packets as necessary. This already happens to a degree.
it's almost 2026 and my isp still doesn't support it
>>107231712There's already been several posts trying to illustrate just how large the address space is. That's the only refutation anyone should need. Of course some people's brains are incapable of dealing with this information and they default to denial/conspiracy/religion mode.
>>107233529>So fucking what?Tech stacks are becoming less efficient by the day, we don't need to make things more complicated when there are existing solutions that just works. This is a solution in search of a problem.>How? This would all be sorted out at the switch/router anyways.How do you subnet a non defined size?>Actually it's 64k. Just that most devices and protocols can't support more than 1500.It's 1500 because of the limitations of ethernet frames, which are 1518 bytes in size. A bigger packet size means higher chances of packet corruption and when a network does not support jumbo frames (more than 1500 MTU) it usually mean packet fragmentation and thus lead to more latency. Almost all networks don't support jumbo frames so you better don't use it except for correctly configured private networks.
>>107215411They'll just do the same things they did with IPv4. That's assuming there isn't an entirely different protocol by then.
>>107234029So ISPs and corporate network folks can just hand out /56s or /48s without fear of them running out?
>>107236463We could literally assign every device its own /64 and we still wouldn't run out any time soon.
>>107234381>Tech stacks are becoming less efficient by the day,Despite what amazon charges, it costs less and less to run stuff every year. Hardware is multiple orders of magnitude more energy efficient and cheaper to produce than it was even a decade ago. We've gotten to the point where embedding 50 different asics into random components on a larger blob of silicone is routine. Look at smartphones. They have hardware accelerated video decoding, GPS calculations, compression algorithms, neural processors for basic AI stuff, etc, etc.>How do you subnet a non defined size?The same way you do normally? You can have subnets for arbitrary bit lengths within the IPv4 space. I don't understand why you think that subnetting dynamically sized ranges would be impossible. An extra step? Yes. Difficult? Hardly.>A bigger packet size means higher chances of packet corruption and when a network does not support jumbo frames (more than 1500 MTU) it usually mean packet fragmentation and thus lead to more latency.Packet corruption is a concern, but we're not talking about doing 64k packets over dialup here. Many datacenters don't even have sub 25Gb networking in them anymore outside of the IPMI network. At gigabit speeds it's about half a second to transfer a full 64kB packet. By the time you get to 100GB you're talking about microsecond intervals. That's still less time tying up a connection than it was 2 decades ago with small packets over a T1, let alone dialup, and you're saving a lot of data.Again, hardware acceleration can reduce this latency down to microsecond intervals. There's zero good reason for equipment on the periphery of backbones to not have this capability and to be capable of negotiating maximum packet sizes, then fragmenting on demand in either direction.Again, I agree that dynamic sizing is stupid, because it's completely unnecessary, but these arguments are silly.
>>107237163>Hardware is multiple orders of magnitude more energy efficientWith "tech stacks" I meant software, not hardware. Protocols do fall under software.>The same way you do normally?It's not the same. You need a maximum size.Variable length addresses implies an infinite amount of them.>you're saving a lot of data.Using 0.061538 % of data (65000 MTU) instead of 2.6667 % of data (1500 MTU).Is it really worth it? 2.6 % difference is negligible. Assuming a perfect internal network connection, if it ever needs to communicate with external networks, it will adds a tons of latency at best or makes the hardware not compatible with said external networks at worst.
>>107233893just call em and ask bro
>>107240883The normal morons at the call center have no idea what IPv6 isAnd if they do, they're still generally not all that informed about it.My ISP for example GENERALLY has IPv6, but only if you're on GPON service, (300mbps up to 1gbps) but if you're on their NGPON2 service (2gbps and occasionally new installs for 300-1gig service) then there is no IPv6 support at all.If I call into my ISP they'll tell me they DO support IPv6 and then try and blame my router settings or PC settings, even though I have 2gig service, and they just don't have IPv6 on the 2gig service.
>>107240911damn yeah it would suck if you couldnt contact the actual engineers directly. total large isp death.
>log onto firewall>go to ipv6>disable itProblem solved
>>107241786what problem? why are you against ipv6?
>>107215411the most annoying thing about ipv6 is it only ended up being used in niche cases. most infrastructures don't support it. there have been many stories of hobbyists trying to switch to it, but they end up getting assraped by the fact that external services only provide limited support.as a consequence, you're forced to learn about all kinds of things ipv6, despite the fact that you are unlikely to ever face it in the wild, and if you do, you will have likely forgotten half of what you learned.
>>107241803Dont need it. I have an ipv4 wan address, all websites I use have ipv4 addressesWhy would I ever have to interact with ipv6?
>>107241848To bring the world closer to a NATless future. Back in the boomer age every device had it's own publicly routable IPv4 address, we must return to tradition. Also, less packet overhead = more speed, less latency.>>107241841>you're forced to learnOH NO THE HORROR!!!
>>107241927yes it is like being forced to take an extra curricular subject like french, something you're unlikely to ever use.not exactly horrifying, but pretty unpleasant.if you can't relate, is it fair to assume you're autistic?
>>107241927That sounds terrible. I dont want jeets to be able to route to my machines. It would be like building a house where every room only had exterior doors. I want my house to have only 1 front door where I can tell you to fuck offAlso what latency? Im sure the latency between the lan port on my computer and the WAN on my firewall is barely measurable
>>107242027Your metaphor is not applicable. Additionally firewalls should drop incoming, untracked connections
>>107242088The whole point of ipv6 is to not have a firewall to decreasse latency / complexityIf you still have one in an ipv6 setup then theres no advantage over ipv4
>>107242115>The whole point of ipv6 is to not have a firewall
>>107242115lol, the point is to not have nat/cgnat.nat is not a firewall.
>>107242115>The whole point of ipv6 is to not have a firewallThe brightest mind of /g/
>>107215411You're probably right in the details you go into but your post viewed from the catalog sounds like you say its useless which it ain't because we need to adress the IPv4 address exhaustion
>>107244816IPv6 bad okay?
>>107215875Epic for the win xD
>>107215411IPv6 is huge. The problem is that windows machines can't be trusted to be "on the border". Everyone who ever read the documentation of CCNA knows that NAT does't mean security, but in reality, it does, at less for windows.Windows was never intended to be used on the internet, it was an OS created by boomers whose idea of "security" was a large nigger with a long list of aggravated assault in his crinal record, standing in front of a door. To make things worse, even the latter added levels of security on windows are so bad that everyone believes it was a deliberately backdoored, left open on purpose.It's so funny you can trust your data on some chinkphone running a sketchy android version directly on the internet, but if you put a windows machine, fully patched, it's a matter of hours before you need to buy monero and transfer to some kid in Romania.
