Table of contents
RIPng (RIP next generation) only supports IPv6. If RIP is wanted for IPv4 as well as IPv6, a RIP process can be run along side RIPng.
It can only be used on small networks (15 or less routers) that aren’t too demanding on network resources.
RIPng is a distance vector protocol. This means that each router believes information received from its neighboring routers about the metric and direction of a particular destination’s path. This is often called “Routing by Rumor”.
RIPng is a simple dynamic routing protocol that’s easy to configure and isn’t very demanding on network resources in the proper use cases. For example, neighbors exchange their complete routing tables every 30 seconds so these tables need to be kept small.
Understanding RIPng is not only important if it is your dynamic routing protocol of choice, but to have a good foundation for learning other dynamic protocols that have additional features and benefits.
How it Works
Adjacent routers can form a neighborship by sharing the same RIPng process. The same network can have a maximum of four RIPng processes running at the same time.
RIPng has two message types:
- Request Messages are sent to neighboring routers to ask for their routing tables.
- Response Messages contain the neighbors' routing table.
Messages are sent to the FF02::9 multicast address using UDP packets on port 521.
- FF02::9 is link-local in scope. The link-local scope is defined by the 2 in FF02. This can be helpful if the global unicast addresses are re-numbered.
- UDP – is a connectionless protocol.
- Port 521 – Using this protocol for in a packet analyzer can be helpful.
RIPng shares Full Updates and Triggered Updates with neighbors so they can update their routing tables:
- Full Updates – Routers send the complete routing table to their neighbors every 30 seconds by default. Routing tables can’t be too large.
- Triggered Updates – If a router learns of a topology change, it will send a copy of its entire routing table
RIPng can do equal cost load balancing for a maximum of four paths per process.
- Update (30 seconds) – Neighbors share a complete routing table.
- Invalid (180 seconds) – If there hasn’t been an update for a prefix by this time interval (180 sec by default), the route is considered invalid and the router will send an infinite metric for this route to its neighbors to poison this route. (Infinite metric and route poisoning are discussed further below).
- Hold-down (180 seconds) – At the time a prefix is considered invalid, the hold-down timer begins. The prefix is suppressed until an update is received with a better metric, the route is flushed or the hold-down timer expires.
- Flush (240 seconds) – begins with invalid timer. When timer expires route is removed from the routing table.
To modify the timers from their default use the timers basic [update invalid hold flush] [value] command.
Administrative Distance and Metric
The AD for RIPng is 120. It isn’t a very “trustworthy” routing protocol. If network resources and expertise of the network administrator allow, a routing protocol with a better AD should be considered. RIPng is better that over-using static routing, even though static routing has a much lower AD. Static routing has a very low AD because it is assumed that the administrator knows what she’s doing.
Administrative Distance is explained with a chart of AD values for commonly used dynamic routing protocols in the IPv6 Routing Table (RIB) lesson.
RIPng (for IPv6) and RIP (for IPv4) use the same metric, but it’s counted differently!
RIP uses hop count as its metric. This is the count of how many hops (links) the traffic must go through to reach its destination. More specifically, a count of one is added to the metric every time the packet reaches the ingress interface of a router. This includes the ingress interface of the receiving device. This results in every link having a count of one. The receiving device will have the sum of all links from the source as its metric for the particular route.
Loopback interfaces are viewed by routers as being Local and Connected as can be seen in the show ipv6 route code. Connected interfaces increase the hop count by one...including loopback interfaces.
Metric is also explained in the IPv6 Routing Table (RIB) lesson.
This is different than IPv4 because IPv4 accepts the metric from its neighbor without adding from its own ingress interface. RIPng will always have a hop count of one higher than RIP for the same topology.
RIPng has a maximum hop count is 15. A route with a metric of 16 means the destination is unreachable (see Poison Reverse below).