Native IPv6

Forums › New Zealand Broadband › Native IPv6

1 | 2

mabnz

43 posts

Geek

ID Verified

#1017204 2-Apr-2014 14:02

I can't believe we're arguing over why dynamic IPv6 is a bad idea when there are 18,446,744,073,709,551,616 addresses in a /64, and the recommended deployment is a /56, which is 256*18,446,744,073,709,551,616.

hashbrown

463 posts

Ultimate Geek

#1017208 2-Apr-2014 14:09

Zeon: That's because you shouldn't be using IP addresses to refer to resources in your network. If you use a name, that can be decoupled from the IP addressing. People who use IP addresses just haven't thought it through and because of NAT - yes it works but is not a good implementation.

All good until you go to write a firewall policy.

michaeln

238 posts

Master Geek

Trusted

#1017222 2-Apr-2014 14:28

mabnz: I can't believe we're arguing over why dynamic IPv6 is a bad idea when there are 18,446,744,073,709,551,616 addresses in a /64, and the recommended deployment is a /56, which is 256*18,446,744,073,709,551,616.

I take it you are arguing that customers should be given a static IPv6 address? (Actually, a static IPv6 prefix allocation). I don't wish to misunderstand you.

1. An IP address is a locator. It is not an identifier. That was also, technically, the case with IPv4, but it was blurred a little in the early days. However, experience led the IPv6 community to quite assertive about this. Address summarisation is essential. It's even more essential with IPv6. There isn't enough memory on the planet to allow for route tables with 2^64 entries.

That being the case, if your location changes, your IPv6 address changes. The subtle point is that your 'location' is defined by your upstream router. So, if you move house, your IPv6 prefix changes; but if your ISP reconfigures its network (changes where in the network your first-hop router is, or changes you over to another first-hop router) then your IPv6 prefix changes. Doing anything else is operationally horrible, expensive, fragile and a Very Bad Idea—no matter how much you might be attracted to a single IP address for life or cafe::babe as part of your prefix.

2. Configuring a server with xxx.yyy.zzz.qqq is OK. Not ideal, but OK. Having to type in xxxx:xxxx:xxxx:xxxx:xxxx:xxxx:xxxx:xxxx is horrible—and error prone. You really want to use dynamic address allocation. You really, really want to use dynamic address allocation for residential connections.

3. Every Internet connected IPv6 device will have multiple addresses. At least 2, possibly more. Just because one of your Internet prefixes changes doesn't mean your LAN stops working.

4. If you are depending on your IPv6 devices having static address, or even static prefixes, STOP. You are doing it wrong. Rethink your design. You don't get to IPv6 best practice by simply rebuilding your IPv4 network with longer addresses.

mabnz

43 posts

Geek

ID Verified

#1017243 2-Apr-2014 14:51

If you give your users IPv6 dynamic addresses (because of the above reasoning) when there is no lack of addresses to go around, you are going to piss a lot of people off! A would imagine the majority of peeps (on Geekzone at least) would appreciate a static address to be able to firewall effectively, and tie in DNS entries (I see you make reference to BIND, but not everyone runs BIND).

michaeln

238 posts

Master Geek

Trusted

#1017274 2-Apr-2014 15:21

mabnz: If you give your users IPv6 dynamic addresses (because of the above reasoning) when there is no lack of addresses to go around, you are going to piss a lot of people off! A would imagine the majority of peeps (on Geekzone at least) would appreciate a static address to be able to firewall effectively, and tie in DNS entries (I see you make reference to BIND, but not everyone runs BIND).

Dynamic DNS is not restricted to BIND. It is commonplace. http://en.wikipedia.org/wiki/Dynamic_DNS

Let's try this differently: if your ISP won't give you a static IPv6 address, how much trouble will it cause you? Not speaking for anyone else—what will you have to do?

mabnz

43 posts

Geek

ID Verified

#1017277 2-Apr-2014 15:24

I'll need to go to the trouble of cancelling my account with them, and I sure as hell won't continue paying them for anything less than static addressing.

Have you ever tried to create firewall policy without knowing the source address of the packet? It's certainly challenging.

hashbrown

463 posts

Ultimate Geek

#1017301 2-Apr-2014 16:04

michaeln: Let's try this differently: if your ISP won't give you a static IPv6 address, how much trouble will it cause you? Not speaking for anyone else—what will you have to do?

How do I open up SSH access to a single host on my network via firewall policy?

Equinox IT

Cloud spending continues to surge globally, but most organisations haven’t made the changes necessary to maximise the value and cost-efficiency benefits of their cloud investments. Download the whitepaper From Overspend to Advantage now.

mabnz

43 posts

Geek

ID Verified

#1017346 2-Apr-2014 16:31

hashbrown:
How do I open up SSH access to a single host on my network via firewall policy?

Precisely.

Beccara

1469 posts

Uber Geek

ID Verified

#1017391 2-Apr-2014 17:31

hashbrown:
michaeln: Let's try this differently: if your ISP won't give you a static IPv6 address, how much trouble will it cause you? Not speaking for anyone else—what will you have to do?

How do I open up SSH access to a single host on my network via firewall policy?

Dont worry the guys all kinds of wrong. There is a massive difference between the ISP handing out static allocations over DHCPv6 vs the ISP handing out random allocations via DHCPv6. On your local lan it's the same, Hard coding v6 isn't really needed much any more (note I said much, there are still valid reasons to do go) but what you can do is assign a static v6 address on your DHCPv6 server so your box will get the same address every time

Most problems are the result of previous solutions...

All comment's I make are my own personal opinion and do not in any way, shape or form reflect the views of current or former employers unless specifically stated

1 | 2