I dont see why Apple is pushing for IPv6 instead of the Network, what are their incentive in doing so? Most Network are now running on Network Grade NAT Ipv4. And many ADSL users as well as users on other countries simply dont have IPv6 equipment.
Edit: I see This is only support IPv6 only, i.e it will still fall back to iPv4. But from experience iPv6 has been causing more headache for users rather then benefits.
3GPP are designing the next generation protocol to replace TCP/IP. And there are other solution such as RINA.
Some are in for a big surprise. I see lots of bad advice being given on the Apple Discussion Forums to turn off IPv6 to resolve home networking issues. Not a good idea since it’s coming sooner rather than later.
There is a few things I don't like about Apples policies but this is one I can easily support. Someone has to set the standards and force the stragglers onto modern protocols. If we don't the risk of running out of addresses becomes very real. So screw the people that can't get with the program!
If there's one thing I think Apple does pretty well, its pushing developers and hardware manufacturers to keep moving forward and eventually stop supporting legacy technology. Sure, it pisses people off (customers and businesses), but I think it makes it hard for Apple to move forward with their technology if they have to keep supporting old technology. In the end, it really only benefits everyone.
I have to disagree the addresses are increased dramatically but that will mean a viable IP solution for years to come. This effectively allows an address for anything on the planet, Mars and anywhere else we go.
Just adding one string to IPv4 would allow for over 100 billion planets, we haven't found a single other planet that we can feasibly go to yet so it would last a while.
Making them human readable is what we have DNS for.
Adding string prefixes opens up all kinds of problems and another gold rush for owning a limited number of language based names. When we start thinking about addressing nanobots and all the other things we will create in the future 832 billion is not nearly enough.
It's over 832 billion billion but it's more about the readability than the number as that can be sorted with any scheme, the alphabet is base-26, alphanumeric is base-36, hex is base-16 so the alphabet is just more concise and readable. The fact that DNS exists is exactly the point, we have to and already do create a human readable alphanumeric mapping to make network addressing usable. Why can't computers just understand the text at the network level in the first place? There would still be a mapping service but it wouldn't be essential. It's the same with telephone numbers. Skype/Facetime don't use numbers, they use strings because that's what humans understand better so that in conversation you can easily tell someone what your contact id is. Landlines still haven't done this even though loads of people have internet. Changing a phone number should be as easy as changing a username.
We only used numbers in the first place because of how slow machines were way back when the internet started out and how connections were made on phone networks and because it was designed by engineers. Switching to hexadecimal (base 16) is nuts, who even knows how to add 1 to a hexadecimal string, it's like having to change HTML hex color codes by having to go into an image program and increment the slider and get the new code. Dealing with sequential address ranges is going to be much more difficult.
Any consecutive groups of zeros can be replaced with :: , saving a fair bit of typing time.
Your thinking is the same as that of the DARPA engineers in the 70's, which has caused our current IP address exhaustion. It's better to have too many addresses than too few.
The scheme I mentioned wouldn't exhaust the addresses, there's a limit with any numbering scheme. It could have been formatted like a fully qualified domain name. Country:Company:Device e.g CN:baidu:255.255.255.255 so every country gets 832 billion billion. You could add a planet code at the front when we find another one. This allows people managing a firewall to easily block an entire country e.g CN:* and exceptions to allow certain companies.
It could be strings all the way down and there can be a localhost id to cover internal usage. You can give devices names directly. An iPhone on an internal network could for example be local:iphone.tim. In the nanobot example, they can be given names like Bob the nanobot and addressed by name. I don't understand the need to deliberately complicate the numbering scheme when we already assign almost every server an alphanumeric domain name, which has to go through DNS resolution every time you connect to it. In the following IPv6 description, you see the word 'confusing' popping up multiple times:
IPv6 has a lot of added advantages so nobody should want to stick with IPv4 but making it confusing to switch isn't helping, they could have taken the opportunity to make it less confusing than IPv4. The new protocol should go even further and have encryption and compression enabled by default. It wouldn't necessarily negate the need for SSL authorities but they wouldn't be a requirement for secure transaction processing or logins. That would get people upgrading right away.
This sort of thing would allow software developers to easily target their own cloud servers without purchasing and configuring domain names. In a game like Candy Crush that needs to handle a lot of data across multiple servers, the company can purchase 10,000 server instances and assign them US:King.com:* where * would be some range of server numbers. In the programming code, they'd be able to directly address those servers without using a domain service and waiting on that propagating or thinking about some odd IP number. They can partition multiple load balanced servers by their function in seconds.
I'm going to call this scheme IPv7. Hopefully it will take off. Like I say, the fact that the IPv6 switch hasn't happened overnight demonstrates that it has been designed badly and you can bet that no matter how much big companies try and push people to switch over, people will still be using IPv4 in 10 years. The following says adoption is at ~10% worldwide in 5 years. It's higher in developed countries but China is around 1%:
Some are in for a big surprise. I see lots of bad advice being given on the Apple Discussion Forums to turn off IPv6 to resolve home networking issues. Not a good idea since it’s coming sooner rather than later.
I just had a Cox Communications guy tell me to shut of iPV6 to resolve some intermittent network outages. Turning it off has not solved them. It will be back on tonight.
Here’s another question. When, if ever (which would mean what’s the point?), will it be possible to force IPv6 connections only to a sufficient amount of services/websites?
So, then doesn't this mean that for 90+% of the wifi routers out there that don't support, or aren't configured to support, IPV6, network-aware apps submitted after July 1 will just stop working until you turn wifi off and use mobile data? Or is there some mysterious built-in workaround I'm not aware of?
If that's right, though, it would also make new apps (spotify, facebook, and a million others) unusable on ipod touch (which can't use mobile data) until you can find an IPV6 wifi network.
So, then doesn't this mean that for 90+% of the wifi routers out there that don't support, or aren't configured to support, IPV6, network-aware apps submitted after July 1 will just stop working until you turn wifi off and use mobile data? Or is there some mysterious built-in workaround I'm not aware of?
If that's right, though, it would also make new apps (spotify, facebook, and a million others) unusable on ipod touch (which can't use mobile data) until you can find an IPV6 wifi network.
It does not mean that.
Apps may still use IPv4, and to not lose functionality on the many networks that don't have IPv6, they probably always will be able to use IPv4. But they must be able to use IPv6 as well. The current practice is to try IPv6 first and to quickly fall back to IPv4 if needed. Apps that already do that shouldn't need any changes to comply with Apple's new policy. But the knowledge that IPv4 is always available as the fallback has allowed developers to just ignore IPv6. No longer. Apple's evaluation of submitted apps will include running in an IPv6-only environment where IPv4 is not available.
Come June 1, all submissions to Apple's App Store must be compatible with IPv6-only standards, the latest internet protocol version for hardware identification and network routing. The company first announced iOS 9 would be transitioning to IPv6-nly network services at last year's Worldwide Developers Conference.
I suspect that the reason for this deadline will be made patently obvious on June 13.
Comments
Edit: I see This is only support IPv6 only, i.e it will still fall back to iPv4. But from experience iPv6 has been causing more headache for users rather then benefits.
3GPP are designing the next generation protocol to replace TCP/IP. And there are other solution such as RINA.
Yes, Apple added IPv6 support 1021 days ago according to the website below. Thats almost 3yrs ago they added support.
http://sixy.ch/
It's over 832 billion billion but it's more about the readability than the number as that can be sorted with any scheme, the alphabet is base-26, alphanumeric is base-36, hex is base-16 so the alphabet is just more concise and readable. The fact that DNS exists is exactly the point, we have to and already do create a human readable alphanumeric mapping to make network addressing usable. Why can't computers just understand the text at the network level in the first place? There would still be a mapping service but it wouldn't be essential. It's the same with telephone numbers. Skype/Facetime don't use numbers, they use strings because that's what humans understand better so that in conversation you can easily tell someone what your contact id is. Landlines still haven't done this even though loads of people have internet. Changing a phone number should be as easy as changing a username.
We only used numbers in the first place because of how slow machines were way back when the internet started out and how connections were made on phone networks and because it was designed by engineers. Switching to hexadecimal (base 16) is nuts, who even knows how to add 1 to a hexadecimal string, it's like having to change HTML hex color codes by having to go into an image program and increment the slider and get the new code. Dealing with sequential address ranges is going to be much more difficult.
The scheme I mentioned wouldn't exhaust the addresses, there's a limit with any numbering scheme. It could have been formatted like a fully qualified domain name. Country:Company:Device e.g CN:baidu:255.255.255.255 so every country gets 832 billion billion. You could add a planet code at the front when we find another one. This allows people managing a firewall to easily block an entire country e.g CN:* and exceptions to allow certain companies.
It could be strings all the way down and there can be a localhost id to cover internal usage. You can give devices names directly. An iPhone on an internal network could for example be local:iphone.tim. In the nanobot example, they can be given names like Bob the nanobot and addressed by name. I don't understand the need to deliberately complicate the numbering scheme when we already assign almost every server an alphanumeric domain name, which has to go through DNS resolution every time you connect to it. In the following IPv6 description, you see the word 'confusing' popping up multiple times:
http://www.techrepublic.com/blog/data-center/breaking-down-an-ipv6-address-what-it-all-means/
IPv6 has a lot of added advantages so nobody should want to stick with IPv4 but making it confusing to switch isn't helping, they could have taken the opportunity to make it less confusing than IPv4. The new protocol should go even further and have encryption and compression enabled by default. It wouldn't necessarily negate the need for SSL authorities but they wouldn't be a requirement for secure transaction processing or logins. That would get people upgrading right away.
This sort of thing would allow software developers to easily target their own cloud servers without purchasing and configuring domain names. In a game like Candy Crush that needs to handle a lot of data across multiple servers, the company can purchase 10,000 server instances and assign them US:King.com:* where * would be some range of server numbers. In the programming code, they'd be able to directly address those servers without using a domain service and waiting on that propagating or thinking about some odd IP number. They can partition multiple load balanced servers by their function in seconds.
I'm going to call this scheme IPv7. Hopefully it will take off. Like I say, the fact that the IPv6 switch hasn't happened overnight demonstrates that it has been designed badly and you can bet that no matter how much big companies try and push people to switch over, people will still be using IPv4 in 10 years. The following says adoption is at ~10% worldwide in 5 years. It's higher in developed countries but China is around 1%:
https://www.google.com/intl/en/ipv6/statistics.html
It has taken years for ISPs to reconfigure their networks because they couldn't trivially map IPv4 onto it.
Apps may still use IPv4, and to not lose functionality on the many networks that don't have IPv6, they probably always will be able to use IPv4. But they must be able to use IPv6 as well. The current practice is to try IPv6 first and to quickly fall back to IPv4 if needed. Apps that already do that shouldn't need any changes to comply with Apple's new policy. But the knowledge that IPv4 is always available as the fallback has allowed developers to just ignore IPv6. No longer. Apple's evaluation of submitted apps will include running in an IPv6-only environment where IPv4 is not available.
I suspect that the reason for this deadline will be made patently obvious on June 13.