National IPv6 launch day in Finland | Hacker News Reader

National IPv6 launch day in Finland

viestintavirasto.fi

166 pointseloycoto108 comments

It bothers me that AWS doesn't have usable IPv6 support (not in VPCs, unavailable for EC2 instances, not S3, missing from Route53, etc).

I think there is a bunch of progressive companies running on AWS that would like to offer IPv6, but cannot.

I agree with you that it is very important that AWS offers it everywhere. However, I would say they do offer it "where it counts": their load balancer a are already dual stack and Route 53 ALIAS records support this directly too. Your clients can access you via IPv6 and that is huge.

But not in S3, or on the CDN, for example.

Their LoadBalancer is IPv6 only in EC2 Classic which is being deprecated.

When you create an AAAA alias record in Route 53 and point it to a VPC load balancer, it gives out an IPv6 address. It does only mention A record in the load balancer description's DNS Name field, so I don't know what to make of it.

Yes, this is very confusing. It assigns the IPv6 address but that does not actually work as the LB is in the VPC which does not support IPv6. Lesson I learned: testing whether IPv6 actually works is worthwile.

That's what I was afraid of...

Apparently Apple announced that once iOS 9 is released, all apps will be required to support ipv6. Given the large number of apps that use AWS as a backend, how will this be feasible?

Supporting IPv6 doesn't necessarily mean IPv6 exclusively. Also, there's always tunneling. Hacky, but it might be the stick that is needed to whip people into actually using IPv6.

The app can support IPv6 while not actually using it for communication. This would be a matter of allowing people to enter IPv6 addresses, websites, e-mail addresses, etc; logging, and displaying IPv6 addresses rather than hard-coding IPv4-only approaches, etc. Basically, prepare your app so that if your backend supports IPv6, your app already does.

It would be bigger news if iOS 9 required all apps to support not having access to IPv4.

There's really no point having IPv6 if you still require IPv4. It's a natural middle step, but you're still burning an IPv4 address.

On the contrary, there are advantages in having your apps/clients support ipv6 even if your backend is not ready to offer v6 connectivity:

1) your app/client will work just fine on ipv6-only networks with NAT64 gateways, making for a better experience for your customers (less breakage on such networks) and will help network operators transitioning to v6,

2) if it does any kind of p2p (think rtp, webrtc, bittorrent, sip, etc.), it will take advantage of nat-free v6 paths for client to client communications, regardless of which address family it uses to talk to the backend.

I doubt that is something that Apple consider to be their problem: what the developer chooses for their back end is their business. If the relevant parts of AWS don't support IPv6 then developers will have two choices:

1. Use some sort of v6-to-v4 proxy service

2. Or chose another system for their backend

Given that Apple are apparently using AWS as part of their iCloud, I suspect it won't be an issue though. Perhaps Apple know that Amazon will be rolling out appropriate changes in a timescale sufficient that they can be reasonably sure there will be no problem?

NAT64 or other transition technologies.

As long as you use DNS instead of hardcoding IPv4 addresses (and if you are using AWS's ELB, you are using DNS instead of hardcoded IPv4 addresses), a mobile phone in an IPv6-only network will connect to your backend as if it had a IPv6 address, with the mobile network doing all the translation.

That's probably why Apple is adding the requirement: no matter whether your backend is on IPv4 or IPv6, your client must be able to connect to it as if it were on IPv6. Due to IPv4 exhaustion and the costs of CGN, the more phones can use exclusively IPv6 (even if through NAT64 or 464XLAT), the better for the network operators.

Using hardcoded IP addresses in your code is a bad idea anyways.

Any of the down voters, do you care to share why you are down voting? This is the official development guide of Amazon as well.

I'm going to hazard a guess and say that Apple probably just wants to make sure that your app doesn't try to do silly things like pass user-input network addresses (like for email apps that need the user to supply an email server address) through parsing regexes like (\d+)\.(\d+)\.(\d+)\.(\d+).

Just for the record this is not the regexp for matching IP addresses. It is for matching 4 multi digit numbers separated by dots.

IP addresses will include IPv6 addresses, so no, this is not a good regex for matching IP addresses. Which is the point.

Even in an IPv4-only world, that regexp admits strings like 123.456.7890.12345.

You're capturing the octets individually. You already have to do bounds checking on RFC 1918 ranges and such. Why wouldn't you do the octet checking in your post-match logic? This is just pedantic.

How about?

^(?:(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.){3}(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)$

Android seems to have been doing a pretty good job at this. I've been using IPv6 exclusively on my Nexus 4 ever since T-Mobile rolled it out, and I've yet to come across an app that isn't compatible with it.

Route53 has IPv6 support in the sense that you can get/set AAAA records. It doesn't have it in the sense that their namesrvers are IPv4 only. That's my biggest annoyance with route53.

(I wrap route53 with tinydns + git to provide DNS with flat pricing and good reliability - http://dns-api.com/ )

Last time I asked a technical(!) sales person at TeliaSonera about IPv6 availability they asked "Did you mean IPTV?".

That pretty much sums it up and they are notably not ready for the "launch", but the launch page implies it's a staged roll out to ensure network stability. I read that as "we couldn't be bothered to get things done in time like the other ISPs". Notice how e.g. DNA just flipped a switch and went bang.

You might be in luck, it seems like Sonera has been deploying 6rd for some time already. Maybe you could already get IPv6?

I have a tunnel already, had for a decade. I want native IPv6.

> technical(!)

Anyone remotely technical knows about IPv6. If that's what you get when you ask for tech support, ask if they have real technicians whom you could talk to.

yes, that person transferred me and then one or two more transfers later I had some actual tech person who went "oh yeah, we do IPv6, I think, it's a business feature. How you get it, I have no clue." It took another two months of probing and poking and unanswered call requests to B2B sales, to find out that in theory I could get a business connection, but it would cost me 10x as much with the only benefit being native IPv6. (It would be still over a crappy VDSL2 line at identical parameters).

> Notice how e.g. DNA just flipped a switch and went bang.

How loud bang and how much smoke?

Purely for the record: http://ape3000.com/ipv6/ - a site that tracks IPv6 deployment in Finland was updated and the "will deploy native for launch" has disappeared next to Sonera and now it says "A schedule for native deployment is not known". So yeah, one of the largest ISPs decided not to give the slightest f. and will just stick to a half-arsed 6rd deployment, it seems.

I think we should pick a day, not too far out, and we all should immediately switch to IPv6 and the metric system world wide. Just rip off the band-aids and move forward.

I can't begin to imagine what that would cost in lost business due to downtime.

This would probably have health implications.

220lb seem more than 100kg

I mean, it could have the opposite affect, since breaking the 3 digit barrier is a bit more important than 220.

kilograms aren't a measure of weight.

You are correct, but when I'm messing around with bumper plates, the red plate says 55kg rather than 539N. If you're not in the lab, a kilogram is simultaneously the mass and the force of gravity of said mass at sea level.

Is "Pound" more like "Newton" than "Gram"?

Correct.

  Pound -> Newton
  Slug[1] -> Gram

[1]: http://en.wikipedia.org/wiki/Slug_%28mass%29

I've never heard of a slug in my life.

Pound by itself usually means mass, but it's interchangeable. Pound-force is a common term. Pound-mass can disambiguate.

Yes, that is what is most common these days. Unfortunately the "slug" unit never caught on, it's way more fun to say than "pound-mass".

They are, in day-to day usage, up to a factor.

> kilograms aren't a measure of weight.

When making obnoxious technical corrections in a discussion, there are two things you must do if you wish to avoid simply being a jerk.

1. Your statement must in fact be technically correct.

2. Your statement must be relevant.

You have failed on #2. The person you responded to said:

   This would probably have health implications.

   220lb seem more than 100kg

Note that nowhere in there did he say anything about "weight". You are simply assuming that he meant "220lb weight seems more than 100kg weight". He could have also meant "200lb mass seems more than 100kg mass".

You have also arguably failed in #1. Kg isn't a measure of weight in the SI system. It is a measure of weight in the de facto system people use in the real world, and is defined as the weight of a 1 Kg mass in a 1 g field. Even scientists use this de facto system much of the time.

> kilograms aren't a measure of weight.

True, but irrelevant to the comparison of pounds to kg since pounds are both a unit of force (sometimes disambiguated as pound-force) and a unit of mass (sometimes disambiguated as pound-mass), with the former equal to the force due to gravity on an object with the mass of the latter in a 1g field.

So, kilograms are directly comparable to pounds (in fact, the pound is defined to be 0.45359237 kg -- and previously, in the US, the kg was defined in terms of the pound as 2.20462 lb -- so its clear that they are measuring the same thing.)

Better pedants, please.

That's why we'd choose the day in advance. Businesses already lose loads of money due to metric<->imperial conversions. Let's stop the madness.

If you want to talk about vast loss of efficiency, let's talk about Europe having three dozen major languages, and ~50% of Europeans sharing no common language. Language is even more important than measurement. One of the biggest competitive advantages the US market has over the EU, is a dominant language.

Obviously Europeans should all convert to either English, Spanish or Mandarin. Now, trying selling that pitch, and the only thing you'll get back is an emotional response about culture.

For better or worse, we already pretty much did convert to English, thank you.

- Northern Europe

The metric system has already been adopted worldwide.

The metric system is a "worse is better" approach. Why is Kg, a base unit, even if it has prefix?

Time in the SI is based on seconds, that relate very odd to the duration of a day.

The same horrible base60 system is used for angles.

Don't use SI. Use a system that fits for your Problem (e.g. the cgs system) Use Units that make sense (like Angstroms or barns) for the problem you are solving. Often SI fits. But SI shouldn't be a dogma.

Don't go fooling yourself into thinking all systems are equally complex just because all systems have sone complexities.

I've read a at least half a dozen studies linking the US standards system to worse learning outcomes. Nobody's saying SI is perfect. It's just a lot, lot less worse.

Of course the US or the Imperial System are a way worse, I never meant to imply otherwise.

But SI is not particular good and it's not a non-brainer to choose SI. Units should be used because they are elegant to use. Like Angstrom and not because they are in the SI system.

And there are other Systems that are just as valid. Instead of SI-Prefixes one could use Japanese numbers. (They have names for the numbers 10^n for n from -24 to 68 in steps of 4.)

Strictly speaking, Kg is a "kelvin gram".

Kilogram (kg) is especially handy, because it's practically same as 1 liter of water, which is 10 cm^3.

It's indeed rather difficult to calculate duration of time. Quick, how many minutes between 6h:37m and 19h:12m?

Angles are simple if you just use radians.

US is very much behind with it. You barely can see road signs with km/h, and temperatures in Celsius. I'd also prefer time to be in HH:MM:SS instead of AM/PM.

https://en.wikipedia.org/wiki/Metrication_in_the_United_Stat...

With this rate it will take forever. There isn't enough political will do finish the job.

Metrification is fine for engineers, but what problem does it solve for people driving cars? Miles are miles everywhere in the United States, and it causes noone confusion. Again, what problem would switching from Fahrenheit to Celsius solve for most people? Rough estimates of outside temperature ranges would have a lower resolution? 100 degrees would be the boiling point of water instead of roughly the temperature of the human body? It would be easier to record measurements when dividing things by 10, but more difficult when dividing by 4 or 8, when in day-to-day life people more commonly divide things into 4 or 8?

Imperial units have the advantage of being human-scale, and consequently they are very useful for when humans want to estimate a quantity. And this is something people want to do all the time.

Europe had the problem of a thousand-plus years of poorly-defined standards that varied from country to country and even within countries. Metrification solved a real commercial problem, and there were so many competing standards that another one didn't really create a problem. This is not a problem the United States has, and metrification proponents always seem to imagine that metrification would be costless.

It is a problem in US. Simply too much inertia and lack of political will to move this forward. Uniformity with other countries helps too. Why do you think military gets it better than others?

0-100 for freezing and boiling points is way simpler to use, especially when 0 is a border for significant change in environmental conditions when temperature goes in either direction from it. Most common temperatures will be two digits.

You didn't actually answer the question of what problem it solves, you just stated that it is a problem.

I did. Simplicity of usage and uniformity.

But how are those problems to drivers today? The vast majority of drivers don't drive between multiple systems very often, so within their experience it is uniform. And what is simpler about km vs miles, given that everyone already knows miles?

I'm not sure why you single out drivers. It's a global change. I just brought examples above where there is an apparent lack of progress in metrification effort. Road signs aren't the only such example.

Just providing that as an example. The same points hold true for temperature. Why is it more simple and more uniform from the perspective of a society that is mostly non metric? What problems would switching solve that people actually deal with on a semi-regular basis? I ask this because unless there are such problems what is the motivation to switch? Why should we?

Those questions should be answered as a whole. And metric systems is simpler overall - there is no debate about it. I don't think it makes any sense to switch to it only partially. The only reason US doesn't switch is inertia (i.e. habit), and not concerns like "how it can improve things".

>The only reason US doesn't switch is inertia (i.e. habit), and not concerns like "how it can improve things"

You say that like they're two different things. No one does anything just because. The benefits of switching have to outweigh the benefits of staying the same, one of which is always going to be "it's already done". Why bother reversing inertia and switching unless it improves things? You aren't articulating why the change should be made, you're just asserting that it should.

And just saying "It's simpler" means nothing, because being simple isn't a benefit all on it's own. Does it being simpler mean X number of increased work hours because we aren't wasting time converting? Does it being simpler mean Y fewer accidents on the road or in the workplace? These are reasons, not just "it's simpler".

You can find details about switching and views on it here: https://en.wikipedia.org/wiki/Metrication_in_the_United_Stat...

US was pushing the effort in the past, with government obviously convinced that benefits are worth it. But inertia was too much, and they ditched active pushing in order not to risk their public image. Since then it's kind of formally "on the plate", but in practice it's barely moving anywhere.

Yes, I understand all that, but it's still not an answer to the question of what, given the facts as they are, the benefits are that make it worth going against that inertia.

(Also: The only discussion on potential benefits in your link is sourced by a dead link)

The benefits are pretty obvious to me - uniformity. I don't think you need any links for it.

I already addressed why this is not sufficient:

>And just saying "It's simpler" means nothing, because being simple isn't a benefit all on it's own. Does it being simpler mean X number of increased work hours because we aren't wasting time converting? Does it being simpler mean Y fewer accidents on the road or in the workplace? These are reasons, not just "it's simpler".

What are the tangible benefits to people's lives that "uniformity" would bring? And do you think they outweigh the benefits of continuing to use our existing system? Why or why not?

Metrification certainly solves a problem for the military, that involves NATO collaboration and international operations. The military already spends a lot of energy training soldiers, so the costs of metrification are lessened. This is not true of most Americans in their day-to-day life.

> This is not true of most Americans in their day-to-day life.

Not really. There is a lot of interactions and Internet age doesn't make keeping this discrepancy easier.

The thing is, these benefits are very small to most Americans while the costs of metrification are pretty large. Most Americans need to drive to an airport and spent half a day or more flying in order to get to a region where metrification is standard. Imperial is used for thousands of miles in every direction, except for specialized use cases.

In most day-to-day life, people are talking about temperature in terms of weather, not freezing and boiling. Weather in Fahrenheit ranges from about 0 to 120 most places in the US most of the time, making it easy to talk about temperature ranges in decades, whereas Celsius ranges from -10 to 45. There is an advantage here either way?

For me it's pretty intuitive, that -1°C and +1°C mean pretty different conditions outside. Anything with - is "cold"/"freezing", anything with + is "warm"/"hot". On the other hand +31°F and +33°F aren't as expressive for the same idea.

Boiling is also a pretty important point (for cooking). So having it at 100 is handy. Water is common in everyday life. Brine? No idea why anyone would use that for common measurements.

Plus, using two digits in the majority of cases is simpler. The bottom line, systems like Fahrenheit end up being used when scientists don't think much about practicality in everyday life. It's good no one is using Kelvin for that purpose.

> -1°C and +1°C mean pretty different conditions outside

They really don't, though. Temperature varies over an area, varies by hour, melting and freezing can take days when you're close to the limit...

It can snow when the surface is above freezing. It can rain when the surface is below freezing.

I can show you two days in any winter, -1C and 1C, and they'll look exactly the same.

And while 100 is convenient for boiling, 200 is roughly just as convenient.

> They really don't, though.

They roughly do. It's just a mental marker (i.e. 0°), which I find to be nicer than 32°.

The whole it just makes more sense for humans argument falls apart when you realize that billions of people do just fine with the metric system.

The whole "[SI] just makes more sense" argument just falls apart when you realize that hundreds of millions of people do just fine with the imperial system. ;)

Hm. Let's see: https://en.wikipedia.org/wiki/Metrication#/media/File:Metric...

I surely see a value in uniform system here. Looks like most agree with me. The only reason not to do it is "I'm not used to it" argument.

Not if you have a technical job imperial is much harder that SI units and remember there is more than one imperial standard eg a Pint in the USA is not a the same as a Pint in the UK

When do you see people dividing things by 8? 10's a lot more useful.

Can you be specific about what you mean by human scale?

Both temperature systems are in the same range. A yard and a meter are almost the same. A mile and a kilometer are in the same range. A pound and a kilogram are in the same range. Cup and liter are both about a factor of two away from a typical glass size. 5mL is about as easy as 'teaspoon' while being clearer and easier to divide into typical minimum recipe amounts.

Where does metric do worse?

I use the DNS flusher addon in Firefox, and with the Status-4-evar toolbar addon, I can see the IP address of the sites I visit. My provider (not in Finland) offers IPv6 support, and I see IPv6 addresses quite often, especially for Google, Facebook and other big sites.

This addon makes IPv6 real for me. I know my provider uses IPv6, but actually seeing this address makes a difference to me. Most people don't know and don't care, and they shouldn't as normal user, but for me it's different.

I like the idea that if you participate you have to leave ipv6 turned on after. These one off days before have just ended and most people turned it off.

Github please add ipv6 support...

And HN too:

    $ host news.ycombinator.com
    news.ycombinator.com is an alias for news.ycombinator.com.cdn.cloudflare.net.
    news.ycombinator.com.cdn.cloudflare.net has address 198.41.191.47
    news.ycombinator.com.cdn.cloudflare.net has address 198.41.190.47

Yes thats odd as you can just turn on ipv6 support with cloudflare.

The only holdback nowadays is typically your application code not being ready to deal with IPv6. It is possible that HN doesn't yet handle features such as no procrastination for IPv6 addresses.

Cloudflare even have a "pseudo IPv4" option for dumb software https://support.cloudflare.com/hc/en-us/articles/202494830-P...

noprocrast is done through user/cookies, I have it enabled but will sometimes switch into a logged out browser to continue reading.

Do you have a paid github account? Please ask for ipv6 support https://github.com/contact

Err, you don't need a paid account to write into Support and ask for a feature. I'm not sure why you mentioned that.

It carries more weight.

I some what skeptical about all those IPv6 launch days. Residential IPSs across the world really do not care. Meanwhile in Canada for example Rogers gives 3G/LTE users IPs from the 25.0.0.0/8 (UK Ministry of Defence) and then NAT them to the pool of the real IPv4 IPs. No IPv6 as you may guess.

As far as I can tell, T-Mobile in the US provisions handsets with IPv6 by default.

How is your experience? If you tether, do you get IPv6 on your laptop as well? Is IPv4 traffic NATed or not ?

I'm not the OP, but when I tether with my Nexus S (running Cyanogenmod), I get an IPv4 address in RFC 1918 space. T-Mobile gives their handsets only an IPv6 address. They also used to use -I think- NAT64, which meant that direct access to resources by their IPv4 address failed. [0] IIRC, their IPv6->IPv4 bridging method has been changed, so you can access IPv4 resources by their address now.

[0] T-Mo was using -I think- the return of an A record in DNS lookup to establish a NAT mapping, so IP access without the lookup didn't trigger the mapping.

I just got an iPhone 6 Plus from T-Mobile a couple days ago, and I don't seem to be getting an IPv6 address over LTE. Maybe they aren't doing it everywhere yet?

I'm in Western Washington, in the Puget Sound area on the other side of the water from Seattle.

On Android, you can just go to "Settings" -> "Cellular networks" -> "Access point names" and choose whether you want to request an IPv4 or IPv6 address.

As far as I can tell (from looking at https://support.apple.com/en-us/HT201699), iOS doesn't have an equivalent option, so it probably just doesn't support IPv6 over cellular networks yet.

Kabel Deutschland utilizes DS-Lite, effectively making the network IPv6 native. Comcast has been doing quite a bit of interesting work with going native IPv6 internally.

I am a huge proponent of LISP and this is one of the "coolest" uses of LISP I've seen yet: https://www.nanog.org/sites/default/files//meetings/NANOG64/...

Comcast is going for a pure IPv6 core network, they are currently at 60% of their customers getting an IPv6 address of some sort (cable modem or even cable box).

T-Mobile's network is IPv6 as well. Especially for customer devices that support 464XLAT: https://sites.google.com/site/tmoipv6/464xlat

There is a LOT of support for IPv6, and many companies like Facebook are finding that IPv6 actually offers better performance end to end for their users.

Take a look at this NANOG presentation: https://www.youtube.com/watch?v=EfjdOc41g0s

To me, enabling IPv6 actually caused some weird issued[0], and had to turn it off... Seems to be more place that can go wrong by just enabling it...

[0] http://forums.xfinity.com/t5/Home-Networking-Router-WiFi/Net...

That's a great idea. I hope other countries follow suit. We need this sort of arbitrary "catalyst events" to sanction migrations, even just at the psychological and political level.

What would have been better is 20-25 years ago take ipv6 out back and shoot it and tel the IETF to go back to the drawing board and come up with one tat had a better migration plan

Well, at least I don't yet have IPv6 working. Might be because this isn't mandatory and my ISP (largest in Finland) just doesn't care or because I'm using 4G since they can't figure out how to get a working fiber connection to me (it works for few hours and then cuts of for 30-90 minutes).

Here also. It's a Saunalahti/Elisa 3G+ dongle. Nothing.

It looks like the seminar is streaming online: http://areena.yle.fi/tv/suorat/yle-areena/1-2779136

Here is info for Elisa Mobile network IPv6 http://www.elisa.fi/ipv6

At least DNA in Helsinki delivered, this is what I started to see today:

    wlp3s0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
            inet 82.181.xxx.xxx  netmask 255.255.248.0  broadcast 82.181.xxx.255
            inet6 2001:14ba:100:0:4a21:b369:xxxx:xxxx  prefixlen 128  scopeid 0x0<global>

http://test-ipv6.com/ gives a score of 10/10. It looks like they are currently serving only /128 via DHCPv6, though.

You actually get a /56 if you ask for it (dhclient -6 -P -N -v -i em1 or somesuch )

Oh, good to know! That was what NetworkManager auto-configured for me, some day I'll have to take a closer look at getting the /56 set up.

Apple has actually done a really good job of pushing IPv6. For years their devices have supported it from their routers to their laptops. Between Apple and Comcast's recent IPv6 turnup I'm actually on IPv6 by default.

IPv6 works without problems in OS X, but AirPort doesn't support IPv6 over PPPoE (I've had no luck with bridge mode either, even though it theoretically should support that).

I've used a Time Capsule just fine in bridge mode when cabled to a CPE sending out router advertisements. If you use it as a bridge, you might want to make sure the TC firewall is disabled (or at least lets multicasts through).

Torille?

Reply on news.ycombinator.com