From 1c8bd5c0965fa550671dcba49f4268b0e475dfff Mon Sep 17 00:00:00 2001 From: Edward Haas Date: Mon, 4 May 2020 09:15:35 +0300 Subject: [PATCH] post: Static IPv6 playground blog Signed-off-by: Edward Haas --- _posts/2020-05-03-static-ipv6-playground.md | 145 ++++++++++++++++++++ 1 file changed, 145 insertions(+) create mode 100644 _posts/2020-05-03-static-ipv6-playground.md diff --git a/_posts/2020-05-03-static-ipv6-playground.md b/_posts/2020-05-03-static-ipv6-playground.md new file mode 100644 index 0000000..b87e5b7 --- /dev/null +++ b/_posts/2020-05-03-static-ipv6-playground.md @@ -0,0 +1,145 @@ +--- +title: "Static IPv6 Playground" +date: 2020-05-03 +categories: + - blog +tags: + - ipv6 +--- + +This post is all about setting up an IPv6 enviorment for learning and testing +of varios setups and scenarios. + +We will use network namespaces for our setup and proceed to define static +IP addresses. +(we will cover dynamic IP addresses in a following post) + +## Basic Setup +In order to prepare the ground for our IPv6 playground, we will use network +namespaces to simulate different network stacks. +A network namespace is an isolated network stack that includes interfaces, +ip addresses and routes. + +Throught this post, we will use the `ip` command which is part of the +[iproute2](https://wiki.linuxfoundation.org/networking/iproute2) utilities. + +### Namespace creation +For our setup, we will use two namespaces: red & blue. + +Lets create our two namespaces: +``` +sudo ip netns add red +sudo ip netns add blue +``` +Each namespace is created with a loopback interface which requires an explicit +enablement: +``` +sudo ip netns exec red ip link set lo up +sudo ip netns exec blue ip link set lo up +``` + +### L2 connectivity (the veth) +In order to enable connectivity between the two namespaces, we will use a +veth interface. A veth interface type comes always in pairs, anything that +ingress one edge, egress the other edge and vice versa. It provides a L2 local +connectivity between the peers. + +We will create the veth interface at the root namespace and then place each +peer in one of the namespaces. + +Lets create the veth interface (and enable its links): +``` +sudo ip link add veth00 type veth peer name veth10 +sudo ip link set veth00 up +sudo ip link set veth10 up +``` +And place each in the relevant namespace: +``` +sudo ip link set dev veth00 netns red +sudo ip link set dev veth10 netns blue +``` + +At this point, assuming that IPv6 is enabled on the host, each namespace +should show a loopback interface and another veth type interface which +has a [link-local IPv6 address](https://tools.ietf.org/html/rfc4291#page-11). + +Lets check the interfaces and their addresses: +``` +$ sudo ip netns exec red ip addr +1: lo: mtu 65536 state UNKNOWN + link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 + inet 127.0.0.1/8 scope host lo + valid_lft forever preferred_lft forever + inet6 ::1/128 scope host + valid_lft forever preferred_lft forever +181: veth00@if180: mtu 1500 state UP + link/ether 92:3a:37:75:36:40 brd ff:ff:ff:ff:ff:ff link-netnsid 1 + inet6 fe80::903a:37ff:fe75:3640/64 scope link + valid_lft forever preferred_lft forever +``` +``` +$ sudo ip netns exec blue ip addr +1: lo: mtu 65536 state UNKNOWN + link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 + inet 127.0.0.1/8 scope host lo + valid_lft forever preferred_lft forever + inet6 ::1/128 scope host + valid_lft forever preferred_lft forever +180: veth10@if181: mtu 1500 state UP + link/ether 8e:6a:6b:83:93:ea brd ff:ff:ff:ff:ff:ff link-netnsid 0 + inet6 fe80::8c6a:6bff:fe83:93ea/64 scope link + valid_lft forever preferred_lft forever +``` + +## L3 connectivity +With the IPv6 link-local addresses in place, we can already check the +connectivity between the two namespaces. + +Using the infromation gathered in the previous `ip addr` commands, learn +the link-local IPv6 address of each peer and use it in the ping command. + +Note: As the addresses have link-local scope, a zone must be added to the +destination address. +A link-local address has a default subnet of 64 bits with a default network +address of `fe80`. Therefore, on a node with multiple interfaces, an explicit +egress interface needs to be provided in order for the packet to know through +which interface to exit. +See [here](https://tools.ietf.org/html/rfc4007) for more information +about zones. + +Lets run an IPv6 ping: +``` +sudo ip netns exec red ping -6 fe80::%veth00 +``` + +## IPv6 with global scope connectivity +The previous IPv6 link-local addresses may be used to check L3 connectivity +betweem two directly connected peers (i.e. interfaces connected to the same +physical LAN). Routers are required not to forward link-local addresses. + +Therefore, in order to enable IPv6 connectivity beyond the physical LAN, +a global scoped address needs to be defined on the interface. +Such address may be set statically or dynamically. + +### IPv6 Global Static Address +In order to enable connectivity without a router, we need both peers to be +set on the same network subnet, i.e. the network prefix need to be identical +for both peers and the host part needs to be unique. + +We will use a 64 bit network subnet with a network address of `fd00`. +Resulting in the following addresses: +- red: `fd00::11/64` +- blue: `fd00::22/64` + +Lets define a static address for each namespace: +``` +sudo ip netns exec red ip addr add fd00::11/64 dev veth00 +sudo ip netns exec blue ip addr add fd00::22/64 dev veth10 +``` +With this behind us, we can check the connectivity (without the zone part +this time): +``` +sudo ip netns exec red ping -6 fd00::22 +``` + +Next we will expore [dynamic IPv6](../dynamic-ipv6-playground).