As there is an upcoming SwiNOG.... lets throw some people under the bus before they arrive. Or at least allow them time to come up with more excuses.
Some quotes from Swiss ISPs from the Call Your ISP page: https://www.sixxs.net/wiki/Call_Your_ISP_for_IPv6
8<----
"Currently, as demand for IPv6 is very low, we have no plans to introduce IPv6 native.
"No plans to support IPv6 for our private and SoHo clients"
"The plan is to move everyone on DSLite."
"Provider info: IPv6 is "planned" and soon should get a priority status. When that "soon" will be is not yet known."
"They know what IPV6 is, eventually they will provide it"
---->8
Come on folks, it is 2016! IPv6 is *20 years* old...
Even Sky.uk was able to get it working[1].
Oh and note: Dual-stack IPv4 + IPv6, along with a /56 per user.
It is not that hard to get right.... and yeah, you kinda had 20 years already to 'plan' for this....
Greets, Jeroen
[1] https://corporate.sky.com/media-centre/news-page/2016/sky-completes-roll-out...
Hi,
On Thu, Sep 15, 2016 at 12:11:44PM +0200, Jeroen Massar wrote:
Oh and note: Dual-stack IPv4 + IPv6, along with a /56 per user.
What do you want this IPv4 stuff for? That's even, like, 40+ years old.
gert
On 2016-09-15 20:04, Gert Doering wrote:
Hi,
On Thu, Sep 15, 2016 at 12:11:44PM +0200, Jeroen Massar wrote:
Oh and note: Dual-stack IPv4 + IPv6, along with a /56 per user.
What do you want this IPv4 stuff for? That's even, like, 40+ years old.
To access those ISPs that didn't bother to move yet ;)
Greets, Jeroen
|Come on folks, it is 2016! IPv6 is |*20 years* old... But still not matured enough to put on public usage, beside of some design flaw it is in some cases even bad implemented
Maybe the isp/hoster/transit provider ned some teaching how to do it the right way.
Em 15 de setembro de 2016 06:11:44 AMT, Jeroen Massar jeroen@massar.ch escreveu:
As there is an upcoming SwiNOG.... lets throw some people under the bus before they arrive. Or at least allow them time to come up with more excuses.
Some quotes from Swiss ISPs from the Call Your ISP page: https://www.sixxs.net/wiki/Call_Your_ISP_for_IPv6
8<----
"Currently, as demand for IPv6 is very low, we have no plans to introduce IPv6 native.
"No plans to support IPv6 for our private and SoHo clients"
"The plan is to move everyone on DSLite."
"Provider info: IPv6 is "planned" and soon should get a priority status. When that "soon" will be is not yet known."
"They know what IPV6 is, eventually they will provide it"
---->8
Come on folks, it is 2016! IPv6 is *20 years* old...
Even Sky.uk was able to get it working[1].
Oh and note: Dual-stack IPv4 + IPv6, along with a /56 per user.
It is not that hard to get right.... and yeah, you kinda had 20 years already to 'plan' for this....
Greets, Jeroen
[1] https://corporate.sky.com/media-centre/news-page/2016/sky-completes-roll-out...
swinog mailing list swinog@lists.swinog.ch http://lists.swinog.ch/cgi-bin/mailman/listinfo/swinog
On 2016-09-19 23:53, Roger Schmid wrote:
|Come on folks, it is 2016! IPv6 is |*20 years* old... But still not matured enough to put on public usage
According to Google 10% of their traffic is IPv6. Apple requires it for IOS.
How is it not 'mature'?
beside of some design flaw it is in some cases even bad implemented
Need more details.
Maybe the isp/hoster/transit provider ned some teaching how to do it the right way.
Management of companies need to be convinced. Technical folks typically know that they want it, but are not allowed to play with it...
That is not a technical, but a political issue.
Greets, Jeroen
Management of companies need to be convinced. Technical folks typically know that they want it, but are not allowed to play with it...
Oh, I have heard a lot of tech excuses:
* Why should I bother with IPv6? IPv4 works fine! * There could be potential security issues with IPv6, so better not use it! * I don't have much experience with IPv6, so better not use it! * I need to add each rule twice on that firewall, that's too complicated and error prone! * Most of the customer accessing this service don't have IPv6, so why should it be reachable via IPv6?
etc.
-Benoît Panizzon-
Just one .. Dropping MTU handling and point to layer7 should handle that doesnt let you feel strange ? So how could an app handle packet size thru L4 ?
My experience is soma pages ar crawling like a snake .. Some ar not loading complete at all, for me v6 is still not ready to deploy to the masses as at least the mentioned flaw is a show stopper
Em 20 de setembro de 2016 02:13:39 AMT, Jeroen Massar jeroen@massar.ch escreveu:
On 2016-09-19 23:53, Roger Schmid wrote:
|Come on folks, it is 2016! IPv6 is |*20 years* old... But still not matured enough to put on public usage
According to Google 10% of their traffic is IPv6. Apple requires it for IOS.
How is it not 'mature'?
beside of some design flaw it is in some cases even bad implemented
Need more details.
Maybe the isp/hoster/transit provider ned some teaching how to do it
the
right way.
Management of companies need to be convinced. Technical folks typically know that they want it, but are not allowed to play with it...
That is not a technical, but a political issue.
Greets, Jeroen
On 2016-09-20 13:00, Roger Schmid wrote:
Just one .. Dropping MTU handling and point to layer7 should handle that doesnt let you feel strange ? So how could an app handle packet size thru L4 ?
Both IPv4 and IPv6 have this little protocol called ICMP (+ICMPv6) it is very useful and for IPv6 it is mandatory.
Even Google (who force MSS to magic values) and Cloudflare had issues with that too:
https://blog.cloudflare.com/path-mtu-discovery-in-practice/
That does not make IPv6 broken though, that makes people who think they have to filter the wrong things broken.
Misconfigurations is not something a protocol can solve.
My experience is soma pages ar crawling like a snake .. Some ar not loading complete at all, for me v6 is still not ready to deploy to the masses as at least the mentioned flaw is a show stopper
I can find many many sites in IPv4 that are brokenly configured. That does not make IPv4 broken.
That you find weird excuses that are already solved for well over 15 years of deployment (even 6bone as shut down 10 years ago....)
Maybe, as it is 2016, time to actually start deploying!?
Greets, Jeroen
Thats why i mentuoned to train techis and your argunent was its a management problem :)
About mandatory how come then https://en.m.wikipedia.org/wiki/Path_MTU_Discovery .. Sorry to point to Wikipedia .. But been on cell only i dont have the proper doc at hand
Em 20 de setembro de 2016 07:09:10 AMT, Jeroen Massar jeroen@massar.ch escreveu:
On 2016-09-20 13:00, Roger Schmid wrote:
Just one .. Dropping MTU handling and point to layer7 should handle
that
doesnt let you feel strange ? So how could an app handle packet size thru L4 ?
Both IPv4 and IPv6 have this little protocol called ICMP (+ICMPv6) it is very useful and for IPv6 it is mandatory.
Even Google (who force MSS to magic values) and Cloudflare had issues with that too:
https://blog.cloudflare.com/path-mtu-discovery-in-practice/
That does not make IPv6 broken though, that makes people who think they have to filter the wrong things broken.
Misconfigurations is not something a protocol can solve.
My experience is soma pages ar crawling like a snake .. Some ar not loading complete at all, for me v6 is still not ready to deploy to the masses as at least the mentioned flaw is a show stopper
I can find many many sites in IPv4 that are brokenly configured. That does not make IPv4 broken.
That you find weird excuses that are already solved for well over 15 years of deployment (even 6bone as shut down 10 years ago....)
Maybe, as it is 2016, time to actually start deploying!?
Greets, Jeroen
On 2016-09-20 13:43, Roger Schmid wrote:
Thats why i mentuoned to train techis and your argunent was its a management problem :)
Management has to provide the training for their people don't they.
Many techies have been asking their managements for years already, but as they do not want to shell out the money/time as their is no real product or urgency.... they'll nicely end up getting service from expensive consultants ;)
About mandatory how come then https://en.m.wikipedia.org/wiki/Path_MTU_Discovery .. Sorry to point to Wikipedia .. But been on cell only i dont have the proper doc at hand
Since when is Wikipedia an RFC?
Greets, Jeroen
Em 20 de setembro de 2016 07:09:10 AMT, Jeroen Massar jeroen@massar.ch escreveu:
On 2016-09-20 13:00, Roger Schmid wrote: Just one .. Dropping MTU handling and point to layer7 should handle that doesnt let you feel strange ? So how could an app handle packet size thru L4 ? Both IPv4 and IPv6 have this little protocol called ICMP (+ICMPv6) it is very useful and for IPv6 it is mandatory. Even Google (who force MSS to magic values) and Cloudflare had issues with that too: https://blog.cloudflare.com/path-mtu-discovery-in-practice/ That does not make IPv6 broken though, that makes people who think they have to filter the wrong things broken. Misconfigurations is not something a protocol can solve. My experience is soma pages ar crawling like a snake .. Some ar not loading complete at all, for me v6 is still not ready to deploy to the masses as at least the mentioned flaw is a show stopper I can find many many sites in IPv4 that are brokenly configured. That does not make IPv4 broken. That you find weird excuses that are already solved for well over 15 years of deployment (even 6bone as shut down 10 years ago....) Maybe, as it is 2016, time to actually start deploying!? Greets, Jeroen-- Sent from my Android device with K-9 Mail. Please excuse my brevity.
On 20.09.2016 13:00, Roger Schmid wrote:
Just one .. Dropping MTU handling and point to layer7 should handle that doesnt let you feel strange ? So how could an app handle packet size thru L4 ?
Uh what now? In IPv4 you have the DF don't fragment bit in the IP header, if set and exceeds a router link, you get an ICMP message back that says fragmentation needed. Your OS handles all this stuff, your application requests and gets a "pipe" to push data through transparently.
In IPv6 the IP header does no longer contain that bit. It's always implied on and if you hit a router where a packet is too big, you'll get an ICMPv6 packet too big message back. Your OS handles all this stuff, your application requests and gets a "pipe" to push data through transparently.
Whenever you want to meddle with such mechanisms, your OS may or may not provide socket hints to programatically influence parameters of said pipe. This is OS dependent and - per sé - has nothing to do with how well IPv6 runs or does not run as it is absolutely not required.
I've activate IPv6 in my home network in 2011 and have never had any problem with it, neither slowdowns nor slow loading pages. Or rather, I do have some slow loading pages, that are all excessivly loading some flash js and other ad stuff. They do the same thing on v4 only as well and are just as slow there.
I do run my own caching-only DNS server, but if this is what is causing "slow page loads", that too is pretty much not v6's problem.
v6 works for the average and even "pro" end user and has been for many years.
Kind regards
My experience is soma pages ar crawling like a snake .. Some ar not loading complete at all, for me v6 is still not ready to deploy to the masses as at least the mentioned flaw is a show stopper
Em 20 de setembro de 2016 02:13:39 AMT, Jeroen Massar jeroen@massar.ch escreveu:
On 2016-09-19 23:53, Roger Schmid wrote: |Come on folks, it is 2016! IPv6 is |*20 years* old... But still not matured enough to put on public usage According to Google 10% of their traffic is IPv6. Apple requires it for IOS. How is it not 'mature'? beside of some design flaw it is in some cases even bad implemented Need more details. Maybe the isp/hoster/transit provider ned some teaching how to do it the right way. Management of companies need to be convinced. Technical folks typically know that they want it, but are not allowed to play with it... That is not a technical, but a political issue. Greets, Jeroen-- Sent from my Android device with K-9 Mail. Please excuse my brevity.
swinog mailing list swinog@lists.swinog.ch http://lists.swinog.ch/cgi-bin/mailman/listinfo/swinog
On 2016-09-20 14:56, René Gallati wrote: [..]
I've activate IPv6 in my home network in 2011
2011, thus 5 years after 6bone had shut down and 12 years after RIR space has been available. Welcome to IPv6! ;)
/me waves at DE-SPACE-19990812 as well, Gert is on this list likely ;)
[..] neither slowdowns nor slow loading pages.
People who have problems with "slow loading pages" have typically two issues:
- MTU issues -- typically as somebody misconfigured MTU or ICMPv6 Packet Too Big is filtered
- DNS recursor that is broken and drops "AAAA" queries -- AAAA resolution needs to time out before A query is done and TCP connection is made
Both are configuration/software issues. People have had quite a while to figure these out and resolve them.... that was the whole point of 6bone...
But come-on that is such a bad excuse in 2016.
Anybody has a proper excuse? :)
Greets, Jeroen
Hi,
On Tue, Sep 20, 2016 at 03:20:56PM +0200, Jeroen Massar wrote:
On 2016-09-20 14:56, René Gallati wrote: [..]
I've activate IPv6 in my home network in 2011
2011, thus 5 years after 6bone had shut down and 12 years after RIR space has been available. Welcome to IPv6! ;)
/me waves at DE-SPACE-19990812 as well, Gert is on this list likely ;)
... which was allocated about two years after we had our first IPv6 router running... that box was decommissioned about 10 years *ago*... :-)
Anybody has a proper excuse? :)
"I can make much more money by selling multi-stage NAT boxes and consulting services to go with it"!
Gert Doering -- NetMaster
On 2016-09-20 15:29, Gert Doering wrote:
Hi,
On Tue, Sep 20, 2016 at 03:20:56PM +0200, Jeroen Massar wrote:
On 2016-09-20 14:56, René Gallati wrote: [..]
I've activate IPv6 in my home network in 2011
2011, thus 5 years after 6bone had shut down and 12 years after RIR space has been available. Welcome to IPv6! ;)
/me waves at DE-SPACE-19990812 as well, Gert is on this list likely ;)
... which was allocated about two years after we had our first IPv6 router running... that box was decommissioned about 10 years *ago*... :-)
You decommissioned the box *10 years ago* and people still do not have IPv6, hahaha ! :)
Anybody has a proper excuse? :)
"I can make much more money by selling multi-stage NAT boxes and consulting services to go with it"!
Which is ABSOLUTELY true, similarly:
"I can make more money by keeping people on IPv4 and then moving them at the last moment to IPv6 when they are willing to pay for it"
or:
"Lets put all the consumers on DS-Lite and sell the IPv4 addresses to businesses with a heavy heavy profit"...
Greets, Jeroen
On 20.09.2016 15:40, Jeroen Massar wrote:
Anybody has a proper excuse? :)
No I don't have an excuse but interested in the communities (& your) opinion re your challenge...
o DHCPv6 re Android re enterprise, BYOD, PWLAN, etc environments
Is a problem or was a problem / no problem at all?
Cheers JIm
swinog mailing list swinog@lists.swinog.ch http://lists.swinog.ch/cgi-bin/mailman/listinfo/swinog
On 2016-09-20 15:58, Jim Romaguera wrote:
On 20.09.2016 15:40, Jeroen Massar wrote:
Anybody has a proper excuse? :)
No I don't have an excuse but interested in the communities (& your) opinion re your challenge...
o DHCPv6 re Android re enterprise, BYOD, PWLAN, etc environments
Is a problem or was a problem / no problem at all?
One person at Google decided that they do not want to properly want to support IPv6 even though people have been asking for it...
https://code.google.com/p/android/issues/detail?id=32621 http://www.techrepublic.com/article/androids-lack-of-dhcpv6-support-poses-se... etc etc
And even if they finally change their mind: https://developer.android.com/about/dashboards/index.html you will never get the correct version on the devices....
I am still waiting for the new 7.x to pop up on that chart ;) I mean, it is only out for a month now...
Hey look, IOS 10.x is at 25% already: https://david-smith.org/iosversionstats/ while that was released, what, a week ago? :)
Hence, just buy a different device, Android is hopeless...
Greets, Jeroen
.... well CyanogenMod might be useful, but they do not have DHCPv6 either afaik...
Well the point wasn't a personal purchase point.
But taking the viewpoint of "someone" who may be deploying / deciding on the deployment of IPv6 in the enterprise, BYOD, etc environments.
With one in two SmartPhones in CH Android, the advice of telling 50% of your customers / users to "go Apple" may not be the best approach for convincing such a "someone".
Seriously, any proposed solutions / approaches (besides requesting google to change their mind), that have actually been used somewhere, would be welcomed?
Cheers JIm
On 20.09.2016 16:24, Jeroen Massar wrote:
On 2016-09-20 15:58, Jim Romaguera wrote:
On 20.09.2016 15:40, Jeroen Massar wrote:
Anybody has a proper excuse? :)
No I don't have an excuse but interested in the communities (& your) opinion re your challenge...
o DHCPv6 re Android re enterprise, BYOD, PWLAN, etc environments
Is a problem or was a problem / no problem at all?
One person at Google decided that they do not want to properly want to support IPv6 even though people have been asking for it...
https://code.google.com/p/android/issues/detail?id=32621 http://www.techrepublic.com/article/androids-lack-of-dhcpv6-support-poses-se... etc etc
And even if they finally change their mind: https://developer.android.com/about/dashboards/index.html you will never get the correct version on the devices....
I am still waiting for the new 7.x to pop up on that chart ;) I mean, it is only out for a month now...
Hey look, IOS 10.x is at 25% already: https://david-smith.org/iosversionstats/ while that was released, what, a week ago? :)
Hence, just buy a different device, Android is hopeless...
Greets, Jeroen
.... well CyanogenMod might be useful, but they do not have DHCPv6 either afaik...
swinog mailing list swinog@lists.swinog.ch http://lists.swinog.ch/cgi-bin/mailman/listinfo/swinog
On 2016-09-20 17:30, Jim Romaguera wrote:
Well the point wasn't a personal purchase point.
But taking the viewpoint of "someone" who may be deploying / deciding on the deployment of IPv6 in the enterprise, BYOD, etc environments.
Various big universities have already been recommending to not buy Android devices. Those same set have actually a plan to stop allowing Android on their network.
The thing is Android 'security' does not really exist: 7.x has been out now for a month, and it does not even show up on the version dashboard: https://developer.android.com/about/dashboards/index.html and the 6.x is from oct 2015, and only has 18% market share.
Now compare that to https://david-smith.org/iosversionstats/ Indeed, IOS10 is about not even a week old and at 25%....
Noting that few folks will be getting 7.x which apparently has some form of 'security updates' at least.... lets see how that pans out.
For IPv6 though, 7.x is not going to fix the lack of DHCPv6... and all the others will never ever get it unless manually installed.
With one in two SmartPhones in CH Android, the advice of telling 50% of your customers / users to "go Apple" may not be the best approach for convincing such a "someone".
Irrelevant when they are not allowed to join the Wifi. People will change their devices over time.
Android lifetime is very short anyway due to their lack of security (stagefright anyone, yep even on 7.x) and updates and people are catching on to that.
Seriously, any proposed solutions / approaches (besides requesting google to change their mind), that have actually been used somewhere, would be welcomed?
Well, you only need to convince 1 person's mind at Google.... noting that I know many Googlers that hate that that single person is making that decision for whatever personal reason....
You can try manually installing the WIDE dhcpv6 client on your phone and repeating that for every phone.
See the bug thread where various folks point out that that just works: https://code.google.com/p/android/issues/detail?id=32621
Greets, Jeroen
That does not make IPv6 broken though, that makes people who think they have to filter the wrong things broken.
Misconfigurations is not something a protocol can solve.
There's an RFC for that: https://www.ietf.org/rfc/rfc4890.txt Great document, even serves as a good primer for folks who are new to IPv6.
On 2016-09-20 19:40, Gregor Riepl wrote:
That does not make IPv6 broken though, that makes people who think they have to filter the wrong things broken.
Misconfigurations is not something a protocol can solve.
There's an RFC for that: https://www.ietf.org/rfc/rfc4890.txt Great document, even serves as a good primer for folks who are new to IPv6.
Unfortunately the people who misconfigure do not read RFCs, if they did, they would not filter.
They do not read this list either, let alone other resources that they should be reading. Hence... not something one can solve.
Greets, Jeroen
Unfortunately the people who misconfigure do not read RFCs, if they did, they would not filter.
They do not read this list either, let alone other resources that they should be reading. Hence... not something one can solve.
BUT: If you find such a person, you can strongly urge them to read this RFC. ;)
Maybe .. But i do not spend money nor time to do some "missionary work" for what reason ever there is when stuff doesn't work. I've done it for years with spammer friendly hoster which even see the abuse box is a Spam trap, and all the so called number one hoster which never take care about what there client do if they pay the rent. In short if a system start to apply some exploits they are blacklisted after second try without notice... If ipv6 starts to troubling and hits my reputation because of some quirks in transfer .. I stop supporting it when ever possible. Business goes before idealism ..
Em 20 de setembro de 2016 15:52:18 AMT, Gregor Riepl onitake@gmail.com escreveu:
Unfortunately the people who misconfigure do not read RFCs, if they
did,
they would not filter.
They do not read this list either, let alone other resources that
they
should be reading. Hence... not something one can solve.
BUT: If you find such a person, you can strongly urge them to read this RFC. ;)
swinog mailing list swinog@lists.swinog.ch http://lists.swinog.ch/cgi-bin/mailman/listinfo/swinog
-
Benoit Panizzon -
Gert Doering -
Gregor Riepl -
Jeroen Massar -
Jim Romaguera -
René Gallati -
Roger Schmid