SwiftOnSecurityTCP/IP is a social construct
Dimly Lit CornersMaybe it's time to renegotiate for a better network protocol?
(eg: https://reticulum.network )
Jima 
@DLC @SwiftOnSecurity I will maintain that "IPv6 is never going to replace it" is an ahistoric vapid hot take, and nobody (the IPv6 weirdos OR the rabid lPv4 adherents) is going to move on to something new without a more compelling argument than that.
("Ahistoric vapid hot take" as in "you don't know how to read graphs, do you?") #UnpopularOpinion
@jima There is no incentive for the USA to support it, they have their IPv4, they are currently totally focused on themselves and actually hostile towards the rest of the world. They have no reason to go to IPv6 just to be able to connect to countries POTUS considers "asshole counties". Yes, these incentives could someday change… who knows? Even if the USA someday become less isolator, open minded, this will take a lot of time & may not happen before #TheJackpot, #TheReset, #WW3, or #Armageddon
@jima And even switching to IPv6 doesn't solve any of the other broken things about the IP protocol stack: DNS will continue to used to spy & censor, BGP will continue to fail, middle boxes will continue to prevent innovation, lacking proper encryption at IP level will continue to cause privacy problems, the cost of domain names will continue to rise, the cost of IPv4 addresses will rise, rent seeking middle men will use these flaws to #enshitify, authoritarians will use these flaws for evil
@jima and yes, we will continue to use this shit internet protocol, if only became we have already invested so much money into it
And there will not be an IPv7 replacing IPv6 to resolve these issues
But maybe, just maybe, when people come together and like want to build their own not-shit robust, resilient, trustworthy, local neighbourhood network, perhaps they could aim for better?
@DLC You're welcome to stay on IPv4 if you like. The world will finish moving on, sooner or later.
Not only will there be an IPv7, there already is one, and it's older than IPv6: https://datatracker.ietf.org/doc/html/rfc1475 (It will not replace IPv6 though, true.)
RFC 1475: TP/IX: The Next Internet
This memo presents the specification for version 7 of the Internet Protocol, as well as version 7 of the TCP and the user datagram protocol. This memo defines an Experimental Protocol for the Internet community. It does not specify an Internet standard.
@DLC People are absolutely welcome to build whatever new technology their heart desires...but that doesn't mean they'll be able to convince or compel anybody else to do so. The slowness (but steadiness!) of IPv6 adoption, an initiative with buy-in from the IETF and the majority of the networking industry, proves how difficult that is.
CyI mean, you're not "convincing" "anybody" to do anything, when it comes to IPv6. You're "forcing" "rotten crooks" to do something. They're already convinced not to, and you can't change their mind, because they're making bank on the exploitation.
Other initiatives that don't have crooked shysters running protection rackets on the artificial scarcity they love, are adopted much more quickly than IPv6.
CC: @DLC@mstdn.social
Other initiatives that don't have crooked shysters running protection rackets on the artificial scarcity they love, are adopted much more quickly than IPv6.
CC: @DLC@mstdn.social
@jima "You're welcome to stay on IPv4 if you like."
Don't be daft, I clearly don't like either
@DLC While I despise how the default implementations of DoH do nothing more than shift who gets to spy on your DNS queries (Cloudflare & Google instead of your ISP), the technology does support proper privacy, when pointed at an entity you trust. (Who you trust is the better question.)
BGP? That's a whole other can of worms to replace.
Middle boxes? Vague.
Encryption at IP level? Ennnh there's a fine line between what can be encrypted and still be able to route.
kasperdIt would be feasible to encrypt a lot more at the IP layer than current protocols do. Whether such an encrypted IP protocol would ever catch on is another question.
There are two fields in the IP header which routing doesn’t have to depend on and for which encryption could improve privacy. The fields I am thinking of are the next header field and the source IP address.
The destination host of course needs to know the source IP, but intermediate routers do not. Intermediate routers only need the source IP for generating ICMP error messages, and we could work around the need for those.
For example I think we could eliminate the need for intermediate routers generating PTB messages by instead have them truncate packets which exceed the outgoing MTU and set a flag in the header to indicate the packet was truncated. Then the destination host can generate the PTB message.
The destination host can be able to see the additional fields by encrypting them using a public key which is embedded into the IP address. That part of the IP address is not used for routing, so it isn’t necessary to include it in every packet. It would need to be part of the IP address records in DNS as well as in the socket API.
An initial handshake between the communicating hosts can establish a shared secret to use, such that asymmetric cryptography doesn’t have to be used on every packet. Such a handshake can also include logic to mitigate IP spoofing that currently exists in some higher layer protocols. Imagine the benefit if the IP layer would do this and it would no longer be necessary for every UDP based protocol to implement cookies.
So I would say this is definitely doable, but it would be a lot of work to implement, and it’s absolutely not guaranteed that it would catch on.
@DLC "There is no incentive for the USA to support it"
This might be news to network operators in the US, which is sitting at 52.58% - 57.98% #IPv6 adoption.*
"they have their IPv4"
Some companies have lots of IPv4, many don't, and many of both need more. CGNAT equipment is expensive, and gets more expensive, the more throughput you need.
@DLC I can't substantially dispute your other points here, other than to say that "connect[ing] to [other] countries" isn't a major argument for deploying v6.
@DLC And further, the Internet doesn't care what POTUS has to say about other countries. Embargoes aside, they're not beholden to adhere to political talking points; if there's money involved in maintaining connectivity to companies in other countries, they'll do it.
Also, classifying that "the USA" has IPv4 is meaningless. It's not one common pool shared by the country; it's many cutthroat companies, each with their own priorities and incentives.
@jima I guess you are more optimistic than me?
For what it's worth, I hope you are right
@DLC Optimist? Not really. I've seen the good and the bad of what big industry is and isn't willing to commit to IPv6, and I've also seen the realities of what staying the course on IPv4 is costing them.
And despite the hand-wringing over how hard IPv6 is to implement, and how they'll need to replace all of their equipment (lies, unless they're idiots), in the long run, IPv6 is cheaper. (Granted, many companies don't care what's past the next quarter or two.)