This is going to be based on a bit of both practical deployment experience and opinion. Let us start with what is acceptable for BGP announcements. RIRs hand out /48s as the smallest allocation. This is perfectly fine for BGP announcements. If you got a larger allocation, say a /32, you should be announcing that master prefix. This will help curb de-aggregation of routes and keep the IPv6 routing table smaller. Is anyone going to tar and feather you for announcing a few specifics? Most likely not, especially if you are providing some sort of anycasted service. Otherwise, until an RIR starts allocating and handing out something more specific than a /48, put in some global filters to make sure nothing gets leaked like /64s.
When providing someone with an abundance of IPv6 subnets (/48, /56, multiple /64s, etc.), make certain that you aren’t putting them ON-LINK, and instead are routing them to a destination IPv6 address on a downstream router. Either by static routes or accepting and properly filtering BGP advertised routes. If you allocate someone a /48 and configure
2001:db8:1::1/48 on the interface facing their router, you basically do a disservice to them. They would need to hack around with proxy NDP and other headaches, rather than have something straight forward and working in seconds.
As far as I’m aware, only 2 popular providers out there seem incompetent regarding this: OVH and FDC. Either they don’t understand how to set static routes on their routers, or don’t want to learn. Seriously, what is so hard about typing in:
ipv6 route 2001:db8:2::/48 2001:db8:1:2::2
With RA, a host will usually configure their default gateway as the link-local address based on the upstream router’s interface, facing that host. For static IPv6 configurations, the
2001:db8:1:2::/64 network address tends to be reserved for just that purpose, being the network address. Some software out there might not like trying to bind on this address, or perhaps there are some older and possibly poorly designed and deployed networking code that won’t like treating that as a host’s address. My recommendation is just use the first “usable” address out of an allocation as the gateway address for hosts, like
2001:db8:1:2::1/64. However any address will work, as with a /64 you get 18+ quintillion of them to pick from (as long as it doesn’t conflict with the host).
Another thing to consider is using /126 allocations for links between routers. This gives a few IPs to use on the link which can be good for multiple BGP sessions, etc. It will also limit the possibility for a ND table overload by someone trying to hit all possible addresses in a larger range, if the link is configured with something smaller (I’ve oversimplified this explanation).
Sure some players like Sprint will plop down
2600:: as that host’s AAAA record and address. Maybe it is a /127 with
2600::1/127 set as gateway; who knows other than Sprint. Point is, to avoid any potential issues, you are better served treating it as the reserved network address for the allocation.
One of the things learned in my time at previous job, was the interoperability between IPv6 routing protocols and platforms. I ran into a quirk when deploying some Quagga boxes (0.99.x version) talking OSPFv3 with some Brocade routers running NetIron 5.x. The issue was that if MTU wasn’t specifically configured on the Brocade interface facing the Quagga box, OSPFv3 wouldn’t properly negotiate and was pretty much useless. However setting the MTU on the Brocade interface to 1500 fixed this. It is possible that this would also work with jumbo frame MTU settings, however the machines weren’t getting configured or deployed for that. Just something to look out for in the lab or production, if OSPFv3 appears configured correctly everywhere, but won’t establish. Quagga’s BGP had no issue with communicating properly with a Brocade router that didn’t specify MTU on its interface, it was only OSPFv3 that appeared to be affected.
So you have your ASN and just got your PI (Provider-Independent Address Allocation) straight from an RIR, or perhaps a LoA (Letter of Authority) to announce some PA (Provider-Assigned Address Allocation) space from your ISP. You’ll need to start getting that announced out there to your peers, customers and transits. So lets use some documentation prefixes and ASNs and sort out a basic working config. I’ll base the examples on my experience with Brocade NetIron software on XMRs, which can translate over to Quagga or Cisco IOS with a few tweaks.
Assuming you are familiar with BGP4+ with IPv4, IPv6 is not so different or any more complex when getting started. Lets start off with some numbers:
Upstream ISP ASN: 64500
Your ASN: 64501
Specific /126 configured on interfaces out of allocated /64: 2001:db8:0:1::/126
PA allocation to announce: 2001:db8:1234::/48
Start off with making certain that you’ve configured IPv6 on your upstream facing interface, and they’ve configured your side, and you can ping each other over the link. The upstream provider’s configuration can be done as so:
isp#conf t isp#(config)ipv6 prefix-list as64501-ipv6-filter seq 1 permit 2001:db8:1234::/48 isp#(config)router bgp isp#(config-bgp)nei 2001:db8:0:1::2 remote-as 64501 isp#(config-bgp)nei 2001:db8:0:1::2 desc Customer_Name isp#(config-bgp)no nei 2001:db8:0:1::2 activate isp#(config-bgp)add ipv6 uni isp#(config-bgp-af)nei 2001:db8:0:1::2 activate isp#(config-bgp-af)nei 2001:db8:0:1::2 filter-list as64501-ipv6-filter in isp#(config-bgp-af)exit isp#wr mem
So the breakdown of these steps is:
- enter configuration mode on router
- build a filter to restrict what the customer (you) are allowed to announce (seq optional but required for multiple entries in list)
- enter BGP configuration mode
- create a specific session using target/destination IPv6 address and ASN of customer
- optionally add a description of the session perhaps to track who it is more clearly
- you don’t want IPv4 routes going out or coming in over the session, IPv6 routes only
- change address-family for IPv6 specific BGP settings
- activate sending and learning IPv6 routes over the session
- apply the filter for accepting INBOUND routes from you/customer
- exit & then write out the config
On the customer side it is similar but not exact:
you#conf t you#(config)ipv6 prefix-list outbound-ipv6-filter seq 1 permit 2001:db8:1234::/48 you#(config)router bgp you#(config-bgp)nei 2001:db8:0:1::1 remote-as 64500 you#(config-bgp)nei 2001:db8:0:1::1 desc ISP_Name you#(config-bgp)no nei 2001:db8:0:1::1 activate you#(config-bgp)add ipv6 uni you#(config-bgp-af)network 2001:db8:1234::/48 you#(config-bgp-af)nei 2001:db8:0:1::1 activate you#(config-bgp-af)nei 2001:db8:0:1::1 filter-list outbound-ipv6-filter out you#(config-bgp-af)exit you#wr mem
The differences being:
1) outbound filter list on your session to the ISP
2) network statement for the allocation to be announced by BGP
You’ll also need some sort of anchor route so BGP knows to announce the route. This can be either a local null-route or static-route for the covering prefix, or an IP out of the prefix configured on an interface. So once BGP is configured, and establishes between both routers, the ISP side should see something similar to:
isp#sh ipv6 bgp nei 2001:db8:0:1::2 routes There are 1 accepted routes from neighbor 2001:db8:0:1::2 Searching for matching routes, use ^C to quit... Status A:AGGREGATE B:BEST b:NOT-INSTALLED-BEST C:CONFED_EBGP D:DAMPED E:EBGP H:HISTORY I:IBGP L:LOCAL M:MULTIPATH m:NOT-INSTALLED-MULTIPATH S:SUPPRESSED F:FILTERED s:STALE Prefix Next Hop Metric LocPrf Weight Status 1 2001:db8:1234::/48 2001:db8:0:1::2 1 140 0 E AS_PATH: 64501
And that should be it. You can obviously play around with peer-groups, route-maps, etc. Communities should work as long as your ISP offers them, so be sure to ask. Ideally there is at least a blackhole community in place. That should allow you to announce a specific range or IP that you want to have null-routed upstream in cases of abuse or attacks. That filter could look like either of the following, depending on how specific they allow you to announce:
ipv6 prefix-list as64501-ipv6-filter seq 1 permit 2001:db8:1234::/48 le 64
ipv6 prefix-list as64501-ipv6-filter seq 1 permit 2001:db8:1234::/48 le 128
With the obvious requirement that the BGP sessions would need to be configured as blackhole community enabled on both sides.
OSS projects and authors are just making it too easy to enable IPv6! Not that this is a bad thing mind you, because it means there is even less reason to NOT have your services available on IPv6. We’ll take a speedy look at configuring Dovecot.
#listen = *, ::
This line lets you configure what addresses Dovecot will listen on.
* = All IPv4
:: = All IPv6
If you have specific IPs you only want Dovecot listening on, you would specify a
listen = line with comma delimited addresses for those IPs. By default Dovecot will listen on all available IPs configured on the server, as will the services you’ve decided to provide (POP3, POP3S, IMAP, IMAPS, etc.).
One of the books I found incredibly useful for becoming more familiar with the basic concepts of IPv6 is Running IPv6 by Iljitsch van Beijnum. It does quite well in explaining key differences between IPv6 and IPv4, and more importantly the benefits of IPv6. The author covers a decent selection of Operating System like: Microsoft Windows, MacOSX, Linux, and various BSDs. As well as Cisco router examples for deploying IPv6 on your network. While it may be circa 2005, the overall basics of IPv6 that are explained in this book are still valid. This is one of the books I’ve recommended over the years, and made certain was stocked on shelves in HE.NET facilities for tech support engineers to have available to read and learn from.
One of the ISP participants in WorldIPv6Launch day is AT&T. They are working towards a goal of having at least 1% of their wired residential customers have IPv6 connectivity. A recent post on the NANOG mailing list states that they have started this with 6rd.
6rd was first launched into production by a residential ISP in France, Free.fr. Comcast was also looking at 6rd for their deployment as part of their trials. Their 6rd trial ended June 2011, and they have since decided that they will deploy native dual-stack to their customers.
Similar to 6to4, 6rd can work for getting your users connectivity, but does require the CPE or directly connected computers to understand and be compatible with 6rd. Another issue would be overloaded relays or what happens when one goes down. Users of 6to4 see this frequently, however with 6to4 it can be a bit more the troubleshoot which relay is having the issue. With 6rd it should be more obvious, and thus quicker to address since the relay is only operated on the residential ISP’s network.