SLAAC versus DHCPv6
-
@Patch said in SLAAC versus DHCPv6:
You are describing the difference between a static and dynamic IP address.
If the ISP honours the DUID the prefix should be essentially static. I've had the same prefix for well over 5 years. That's "static" enough for me.
-
@JKnott
Not sure why that was addressed to me. I was describing the different use case for static and dynamic IP, not how to get a static IP. -
Well, I did respond to your post. However, the last line about dynamic is better for privacy is wrong. With SLAAC you get up to 7 privacy addresses, based on a random number. You get a new one every day, with the oldest falling off the list. The most recent is the address used for outgoing connections. The consistent address, which can be based on either the MAC address or a random number, can be used for things like VPN access to your network. Unless the DHCPv6 address changes every day, it's actually worse for privacy.
-
@JKnott I don't know if you have Comcast or not. I am fighting them for a site I just hooked up in Northbrook, Illinois. Every 24 hours they re-allocate the darn IPv6. They are handing out a /60 but it's a PITA I will be removing their Xfinity gateway and going with a vanilla cable modem and seeing if this gets rid of the IP address swaps they are doing. Comcast plays so many games to upsell features you will never use though their crappy gateway to get a discount.
-
I'm on Rogers and they use the same equipment as Comcast. However, there is a setting in pfSense that may affect this. It's System / Advanced / Networking / Do not allow PD/Address release. If that's not selected, the prefix will change frequently.
-
@JKnott said in SLAAC versus DHCPv6:
line about dynamic is better for privacy is wrong. With SLAAC you get up to 7 privacy addresses, based on a random number.
they are all in the same range provided by the ISP, readily revealed by masking the address of any one of these addresses.
- A dynamic ISP address range publishes which ISP the user is connected to the internet by
- A static ISP address publishes your personal address range
Then if one of the devices on your network has location services enabled for "Network & Wireless" or through your browser then the "privacy" addresses have your street address encoded in every internet communication.
Recording addresses over time is likely to reveal what the 7 "private" addresses are for each device.
So a static IP address range is not really very private at all in my opinion.
-
@Patch said in SLAAC versus DHCPv6:
they are all in the same range provided by the ISP, readily revealed by masking the address of any one of these addresses.
A dynamic ISP address range publishes which ISP the user is connected to the internet by
A static ISP address publishes your personal address range
recording addresses over time is likely to reveal what the 7 "private" addresses are for each deviceMy WAN address is provided by DHCPv6, along with my prefix. It makes no difference whether I use SLAAC or DHCPv6 on my LAN, with regard to my prefix. It will be the same either way. The only issue is whether the ISP will honour the DUID.
One other thing, with the huge address block within a /64, an attacker would have a hard time finding something to attack, even if they knew the prefix. A single /64 contains 18.4 billion, billion addresses. Compare this to a bit over 4 billion for the entire IPv4 address range. Of course there's a firewall called pfSense that goes a long way to keeping attackers out!
-
@JKnott said in SLAAC versus DHCPv6:
One other thing, with the huge address block within a /64, an attacker would have a hard time finding something to attack, even if they knew the prefix. A single /64 contains 18.4 billion, billion addresses. Compare this to a bit over 4 billion for the entire IPv4 address range. Of course, there's a firewall called pfSense that goes a long way to keeping attackers out!
Now do the math with a /60: 16 x 18.4 billion - There is a reason I never see IPv6 address scans, but I still see port scans once a nefarious site logs your IPv6 they have the basic IP subnet, and it would take forever to scan through all those addresses.
-
With my /56, it's 256 /64s. Yeah, it would take a while. I mentioned privacy addresses. They have a lifetime of 7 days. After that, they'd have to start over to find another address.
-
Here is the guy who started the conversation. Thank you very much for all the information and the discussion. This helps me.
Conclusion:
Use SLAAC for clients and DHCPv6 for servers.@keyser said in SLAAC versus DHCPv6:
is a PITA to get IPv6 working properly
+1
IPv6 is much more complicated than IPv4. And that's okay. Everybody involved in networking should be able to learn the IPv6 basics.
But the ISPs have messed it up, espceially with frequent prefix renewals which are heavy to handle and unnecessary. Three or four IPv6 prefix renewals per year would be enough. -
@JKnott said in SLAAC versus DHCPv6:
It makes no difference whether I use SLAAC or DHCPv6 on my LAN, with regard to my prefix
Correct
@JKnott said in SLAAC versus DHCPv6:
One other thing, with the huge address block within a /64, an attacker would have a hard time finding something to attack
That is not an address privacy issue, but it maybe part of a network servers exposure to the internet. Network protection by obfuscation is not really a good approach, as it fails repidly if someone takes the time to look especially if they can get a hint were to look from another source (human engineering, traffic monitoring etc). Direct protection using a decent firewall is far better.
@JKnott said in SLAAC versus DHCPv6:
With SLAAC you get up to 7 privacy addresses, based on a random number.
As I have tried to explain, 7 networks addresses is a tiny number and everyone of these addresses has the same prefix.
-
If your ISP is giving you a static prefix, the the prefix will almost certainly encode / reveal your street address every time you make an internet connection from any device on your local network. My interpretation of which is the "SLAAC Privacy addresses" you are using maybe making you feel you have achieved something but they actually provide almost no privacy functionality.
-
If your ISP is giving you a dynamic prefix, the prefix will encode / reveal what ISP you are using but not which service or your street address. If you then add some randomisation of the lower order bits for each device you may achieved some privacy (not as much as a VPN or routing randomisation Tor tries to achieve).
@Jung-Fernmelder said in SLAAC versus DHCPv6:
ISPs have messed it up, espceially with frequent prefix renewals which are heavy to handle and unnecessary.
They are offering some internet privacy to those user who like some privacy. If you don't value privacy then perusing the options for a static IP address would be appropriate for you.
-
-
@keyser said in SLAAC versus DHCPv6:
is a PITA to get IPv6 working properly
Well could you, or someone else, explain for non native english speakers what "PITA" in this context means. If i g00gle it i just learn what i already know, PITA is nice greek food. But for sure you don't talk about nice greek food but more about something like a pain ;-)
-
@eagle61 said in SLAAC versus DHCPv6:
@keyser said in SLAAC versus DHCPv6:
is a PITA to get IPv6 working properly
Well could you, or someone else, explain for non native english speakers what "PITA" in this context means. If i g00gle it i just learn what i already know, PITA is nice greek food. But for sure you don't talk about nice greek food but more about something like a pain ;-)
PITA = Pain In The As* -
@keyser said in SLAAC versus DHCPv6:
PITA = Pain In The As*
Thanks very much. And yes food also may some times result in PITA, but more likely not greek food, but maybe very spicy food from India or other south east Asia countries ;-)
I know its off topic - sorry for that.
-
I also had to google the abbreviation PITA.
@Patch said in SLAAC versus DHCPv6:
They are offering some internet privacy to those user who like some privacy.
It makes sense that frequent changing prefixes improves privacy. A static or seldom changing prefix is like a static or seldom changing IPv4 adress - anyone can recognize somebody as a user of a specific line. Since typically not more than five ´people share one private internet access subscription there are few doubts about the users identity.
-
@Jung-Fernmelder said in SLAAC versus DHCPv6:
IPv6 is much more complicated than IPv4. And that's okay. Everybody involved in networking should be able to learn the IPv6 basics.
Well, I've been running IPv6 on my home network for over 14 years. Haven't had a problem with it. It works fine. Most of the basics are the same or similar to IPv4. However, there are some significant changes in things like ICMP, no more ARP or broadcasts etc..
-
Unless an ISP provides a consistent host name, changing the prefix will make it impossible to have a VPN to that network, just like with IPv4.
Here's a comparison. My ISP provides a /56 prefix and both IPv4 and IPv6 hostnames are tied to the modem and firewall MAC addresses. Even if my addresses change, I can still use the host name. On the other hand, a friend is on Bell Canada. First off, they don't provide IPv6 at all. Also, the IPv4 address changes frequently and the host name with it. This means to access a server he has on his network, he has to use dynamic DNS and that sometimes fails. Which would you rather have? Also, if I want to have my prefix change, it's a simple matter of a single check box in pfSense to let that happen.
As for servers, if you want others to access them, you have to have a DNS record for it on a public DNS. In this regard, it makes absolutely no difference whether you use SLAAC or DHCPv6 on the local LAN.
-
@JKnott said in SLAAC versus DHCPv6:
As for servers, if you want others to access them, you have to have a DNS record for it on a public DNS.
+1
Applying static IP adresses, both IPv4 and IPv6, combined with normal DNS records is the only professional way to make servers accessible to the public.
But subscribing a static IPv4 adress and a static IPv6 prefix costs money. A VPS hosted in a data center may be cheaper. Or craft a homebrew solution with DynDNS for non-critical scenarios or the private IT playground.@JKnott said in SLAAC versus DHCPv6:
it makes absolutely no difference whether you use SLAAC or DHCPv6 on the local LAN
It makes a different: If you use DHCPv6 only Android clients wont't have IPv6 connectivity because Android doesn't support DHCPv6.
Edit note 2024-09-03 14:12 UTC: Forgotten blank line added.
-
@JKnott said in SLAAC versus DHCPv6:
Unless an ISP provides a consistent host name, changing the prefix will make it impossible to have a VPN to that network, just like with IPv4.
Well my ISP changes prefix every 24 hours. Still i use VPN to that network, just like with IPv4. I use a free of charge DynDNS-Provider to fix that.
-
@Jung-Fernmelder said in SLAAC versus DHCPv6:
Applying static IP adresses, both IPv4 and IPv6, combined with normal DNS records is the only professional way to make servers accessible to the public.
I use a public DNS for my addresses. However, I'm the only one that uses it.
It makes a different: If you use DHCPv6 only Android clients wont't have IPv6 connectivity because Android doesn't support DHCPv6.
Are we talking about the WAN side or LAN. You seem to be confused. On the WAN side there's usually no option, it's DHCPv6. On the LAN side, where you'll find Android clients, you can choose SLAAC or DHCPv6. What's on the WAN side has absolutely no effect on what you use on the LAN side.
