Internet Protocol Version 8 (IPv8)

Gimli_ said:
Dual stack is annoying, agreed that is why we should push for ipv6 transition everywhere. In SA we are at 3% of all traffic in other countries eg. France, Germany, India they are averaging +70%. It is not as if we have s choice to ignore ipv6.
Click to expand...
If it hasn't by now, it never will. Something similar to IPv8 is far more likely if it moves forward.
 
koffiejunkie said:
Something to know: Anyone can submit a proposal to the IETF. The fact that this was proposed, and thus published as a proposal, means absolutely nothing. What's even worse, It adds more complexity without solving any real problem. Most computer users don't ever see an IP address, and if they did, it wouldn't mean anything to them anyway.

Networking professionals have been learning IPv6 as part of their certifications for decades now - CCNA introduced it in 2007. That leaves the rest of us working in IT, who should have skilled up on it at least a decade ago. Linux has supported IPv6 since 1996. The BSDs and Apple followed in 1999/2000 or thereabout. Even Microsoft got it done in 2001 with XP. So there really is no excuse any more.

Calling IPv6 a mess just advertises that you didn't bother learning it. The tricky bits - sub-netting, routing, etc - are hardly any different from v4, it's just bigger numbers and more octets. The rest is just new ways of doing things, which we deal with in every other domain anyway.

I think it's high time for the big tech players - Google, Microsoft, Amazon, Apple, etc - to get together an announce that they'll be discontinuing IPv4 support in 10 years or some such. That will get the last holdouts to get their act together. It's high time for CGNAT to die a miserable death.
Click to expand...
The reality is simply:
  • No business spends millions on a forklift upgrade for zero net ROI.
  • IPv6's fatal flaw is zero backward compatibility.
  • Until a transition doesn't disrupt legacy systems, IPv4 (and CGNAT) isn't going anywhere.
 
quovadis said:
No business spends millions on a forklift upgrade for zero net ROI.
Click to expand...
Business do spend money when their continued existence depends on it. Just like when Google announced they would de-rank non-https sites in their search results, and overnight everyone got on board with SSL certs.

quovadis said:
Until a transition doesn't disrupt legacy systems, IPv4 (and CGNAT) isn't going anywhere.
Click to expand...

Transition options have been around for a long time. Here's one for this use case: https://www.cisco.com/c/en/us/td/do...tion/xe-16-10/nat-xe-16-10-book/iadnat-46.pdf

And let's not forget that a system that's legacy enough to lack IPv6 support shouldn't be anywhere near the internet at all, making this discussion entirely moot.
 
koffiejunkie said:
Business do spend money when their continued existence depends on it. Just like when Google announced they would de-rank non-https sites in their search results, and overnight everyone got on board with SSL certs.
Click to expand...
Search engines and browsers penalised non-HTTPS because it directly affected user safety and trust. IPv6 adoption doesn't solve anything for business or customers of those businesses.
koffiejunkie said:
Transition options have been around for a long time. Here's one for this use case: https://www.cisco.com/c/en/us/td/do...tion/xe-16-10/nat-xe-16-10-book/iadnat-46.pdf
Click to expand...
Backward compatibility is when newer systems can interoperate with older ones without burying the mismatch behind protocol mangling.
koffiejunkie said:
And let's not forget that a system that's legacy enough to lack IPv6 support shouldn't be anywhere near the internet at all, making this discussion entirely moot.
Click to expand...
“let's not forget that a system that's legacy enough to lack IPv6 support shouldn't be anywhere near the internet at all” is nonsense. IPv4-only does not mean insecure, abandoned or legacy. “Moot” assumes the hardware is the problem. The blocker is people.
 
quovadis said:
Search engines and browsers penalised non-HTTPS because it directly affected user safety and trust.
Click to expand...

You're missing the point. Google enforced a changed, everybody else acted. Not because every webside cared about their customers' security, but because they wanted to protect their revenue.

quovadis said:
Backward compatibility is when newer systems can interoperate with older ones without burying the mismatch behind protocol mangling.
Click to expand...

I said nothing about backward compatibility. I don't care about backward compatibility. I have you a transition option in-line with the one you mentioned (CGNAT)

quovadis said:
“let's not forget that a system that's legacy enough to lack IPv6 support shouldn't be anywhere near the internet at all” is nonsense.
Click to expand...

Again missing the point - we're talking about legacy systems. Do you understand what legacy means? Anything that's current and lacks IPv6 is one software update away from having IPv6.

quovadis said:
IPv4-only does not mean insecure, abandoned or legacy.
Completely irrelevant to my comments. You brought up legacy systems - don't try to confuse things.
Click to expand...
Click to expand...
 
Wouldn’t it have been far simpler to just add another octet or even two or three to the IPv4 system?

And in that way retain a measure of backwards compatibility and simply extend on the existing system.
 
koffiejunkie said:
You're missing the point. Google enforced a changed, everybody else acted. Not because every webside cared about their customers' security, but because they wanted to protect their revenue.
Click to expand...
Sure, noone cared about security. :laugh: Hijacking of session identifiers in plain text didn't lose any company any money from fraud.
koffiejunkie said:
I said nothing about backward compatibility. I don't care about backward compatibility. I have you a transition option in-line with the one you mentioned (CGNAT)
Click to expand...
We're discussing ipv8 as a solution which is backward compatible and you were responding to my point in respect of supporting legacy systems...
koffiejunkie said:
Again missing the point - we're talking about legacy systems. Do you understand what legacy means? Anything that's current and lacks IPv6 is one software update away from having IPv6.
Click to expand...
You clearly don't have experience in industries where 'legacy' means a device cannot legally or physically be altered without millions of dollars in recertification and years of clinical validation. Healthcare, aerospace, industrial automation are all industries that have such legacy limitations.
 
koffiejunkie said:
IPv4 was very far from being the dominant standard when work on IPv6 started. Windows didn't even have an IPv4 stack back then.
Click to expand...
Windows 95 had an IPv4 stack natively as did Win NT 3.1 and Windows for Workgroups 3.1 with the add on pack. All predated IPv6's "6bone" launch in 1996 which was really IPv6 encapsulated over an IPv4 tunnel. Production allocations for IPv6 were only in 1999 along with limited commercial-grade backbones. IPv4 very much was the only dominant and only adopted standard. Even prior, Trumpet Winsock served Windows users and Linux's kernel had early support in 94.
 
SauRoNZA said:
Wouldn’t it have been far simpler to just add another octet or even two or three to the IPv4 system?

And in that way retain a measure of backwards compatibility and simply extend on the existing system.
Click to expand...
No, you'd still sit with transition mechanisms, hardware incompatibility, and incompatible protocols. The podcast I posted earlier in the thread has a good discassion about it.
 
quovadis said:
Sure, noone cared about security. :laugh: Hijacking of session identifiers in plain text didn't lose any company any money from fraud.

We're discussing ipv8 as a solution which is backward compatible and you were responding to my point in respect of supporting legacy systems...

You clearly don't have experience in industries where 'legacy' means a device cannot legally or physically be altered without millions of dollars in recertification and years of clinical validation. Healthcare, aerospace, industrial automation are all industries that have such legacy limitations.
Click to expand...
There is no IPv8, and this proposal will at best have the same challenges as ipv6
 
Gimli_ said:
Dual stack is also just a transition mechanism.
Click to expand...
A transition that was estimated to be completed in the mid 2000s if memory serves and at this rate at least another 20 years unless IPv4 is retired which imho would be unlikely.

Gimli_ said:
Make peace with it and teach yourself a new skill and move with the rest op the world.
Click to expand...
My first IPv6 implementation was 20 years ago. I thought we were having a general discussion here...
 
quovadis said:
Sure, noone cared about security. :laugh: Hijacking of session identifiers in plain text didn't lose any company any money from fraud.
Click to expand...

eCommerce sites were literally running X-Cart over plain text because they cared more about the cost of a Thawte-123 certificate than their customer data. Sure, we'll send your card data over SSL to the payment provider, but only after getting it from you over plain text. Did you forget how we got to everything-SSL?

quovadis said:
We're discussing ipv8 as a solution which is backward compatible and you were responding to my point in respect of supporting legacy systems...
Click to expand...

Yes, that's what the article is about, and I am opposed to that idea and in favour of letting IPv4 on the public internet die already. I stand by that. My proposal, to let a few of the big cloud providers stop offering IPv4 addresses, does not mean you can't keep your special snowflake device working. All it will mean is ISPs and internet properties that don't yet have IPv6 will spend the next 10 years enabling it.

quovadis said:
You clearly don't have experience in industries where 'legacy' means a device cannot legally or physically be altered without millions of dollars in recertification and years of clinical validation. Healthcare, aerospace, industrial automation are all industries that have such legacy limitations.
Click to expand...

You should not make assumptions. And as we've established (although apparently it goes over your head) there are ways of keeping

quovadis said:
Windows 95 had an IPv4 stack natively as did Win NT 3.1 and Windows for Workgroups 3.1 with the add on pack. All predated IPv6's "6bone" launch in 1996 which was really IPv6 encapsulated over an IPv4 tunnel. Production allocations for IPv6 were only in 1999 along with limited commercial-grade backbones. IPv4 very much was the only dominant and only adopted standard. Even prior, Trumpet Winsock served Windows users and Linux's kernel had early support in 94.
Click to expand...

I did say "work started" but fair enough.
 
koffiejunkie said:
eCommerce sites were literally running X-Cart over plain text because they cared more about the cost of a Thawte-123 certificate than their customer data. Sure, we'll send your card data over SSL to the payment provider, but only after getting it from you over plain text. Did you forget how we got to everything-SSL?
Click to expand...
Nah, most e-commerce sites encrypted the checkout process for years before the whole everything ssl.
koffiejunkie said:
Yes, that's what the article is about, and I am opposed to that idea and in favour of letting IPv4 on the public internet die already. I stand by that. My proposal, to let a few of the big cloud providers stop offering IPv4 addresses, does not mean you can't keep your special snowflake device working. All it will mean is ISPs and internet properties that don't yet have IPv6 will spend the next 10 years enabling it.
Click to expand...
Or everyone will migrate to the competition that doesn’t axe ipv4 which is probably more likely.
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X