06-25-2019 08:07 AM
Our network uses static routes and VLANs to direct and separate traffic. I do not believe we have a routing protocol enabled. If I enable EIGRP will my network stop working correctly?
Solved! Go to Solution.
06-25-2019 08:26 AM
Hello transaxle,
short answer is no if you are using static routes with default admin distance of 1 ( no floating static routes with final parameter AD greater then 90 or 170 for external EIGRP routes).
However, to have EIGRP to send updates to other routers you may need to redistribute static routes into EIGRP in each node.
With OSPF you don't need to perform this redistribution and OSPF should be able to populate its own link state database even if there are no OSPF routes installed in the routing tables for its link state nature.
Hope to help
Giuseppe
06-25-2019 08:26 AM
Hello transaxle,
short answer is no if you are using static routes with default admin distance of 1 ( no floating static routes with final parameter AD greater then 90 or 170 for external EIGRP routes).
However, to have EIGRP to send updates to other routers you may need to redistribute static routes into EIGRP in each node.
With OSPF you don't need to perform this redistribution and OSPF should be able to populate its own link state database even if there are no OSPF routes installed in the routing tables for its link state nature.
Hope to help
Giuseppe
06-25-2019 10:53 AM
Thanks! That is very helpful. I will learn more about OSPF. Which routing protocol do you prefer?
06-25-2019 12:16 PM - edited 06-25-2019 12:23 PM
One might argue that EIGRP is technically slightly superior to OSPF (although usually it doesn't matter much) but OSPF's major advantage over EIGRP is that's it's not Cisco proprietary. (BTW, my experience has been to use OSPF in large Enterprise networks. Cisco's implementation of OSPF is often superior to other vendor's.)
06-25-2019 12:51 PM
Hello transaxle,
I have more experience in OSPF.
However, depending on your topology and your IPv4 address plan EIGRP can be a good choice too.
For EIGRP to scale and be stable and reliable in a large network you need to limit the Query scope.
EIGRP is the only protocol that involves other neighbors when a network topology change occurs and there is not "acceptable" backup route ready to use.
This EIGRP distributed computation is at the same time a good thing and a problem if too many devices are involved the chances for a so called stuck in active increases.
To minimize Query scope the leaf routers should use the EIGRP stub feature. And appropriate route summarization and route filtering should be implemented on distribution / core routers.
EIGRP biggest advantage over OSPF is the capability to summarize routes out a generic interface.
EIGRP can be a better choice in some specific topologies hub and spokes or services like DMVPN
OSPF to scale needs to use a multi area design with area 0 in the middle and all other areas need to connect to area 0 both for routing exchange and for traffic forwarding. Route summarization can be performed in OSPF only on area border router connecting area 0 and another area. Route summarization at area border can be performed only for internal routes.
In brief EIGRP is easy to configure in a small network but requires additional tuning for scaling.
OSPF requires to study a little more to start using it, but it is a standard based protocol that allows for multi vendor networks and for MPLS services.
Hope to help
Giuseppe
06-25-2019 11:03 AM
We really do not know enough about the environment of the original poster to be able to give detailed answers yet. We can say that in general adding dynamic routing with EIGRP to a network currently using only static routes should not break anything. We can say that if routing is using normal static routes (without specifying administrative distance) and if EIGRP advertises a route for which there is already a static route (for the same prefix and mask) that the router will prefer the static route over the EIGRP route. In that sense not much changes and adding EIGRP will not break anything. But if EIGRP is enabled and if EIGRP advertises a route for which there is not already a static route (for the same prefix and mask and which would have been using the default route) then the EIGRP route will be used and the network forwarding will be different. In general we would expect the new EIGRP route to be an improvement and not break the network. But I can think of scenarios where the EIGRP implementation has some flaws and which could conceivably wind up breaking the network. So you need to evaluate carefully your implementation of EIGRP.
HTH
Rick
06-25-2019 12:21 PM
06-25-2019 12:56 PM
I had some additional thoughts:
Here is more detail about my network and what I want to do with it. I have a core stack of four layer-3 switches and about ten layer-2. Since I don't have more than one layer-3 core, would enabling routing even help me?
06-25-2019 01:07 PM
06-25-2019 01:19 PM
Thanks! Yes, they are “stacked” as one core. I think the static routes I was thinking of are simply the subnets (VLANs). Thanks for suggesting I give more detail. It made me realize I don’t even have two independent layer-3 devices.
06-25-2019 03:56 PM
06-26-2019 07:37 AM
If you have a single routed core stack with multiple connected access switches operating as layer 2 switches then I see no need (and no advantage) in enabling a dynamic routing protocol. I started to mention the possibility of a static default route. That brings up the question of Internet access. And to get Internet access would require address translation. I am guessing that your stacked core switches do not support address translation (you need a quite high end switch to get support for address translation). So do you need Internet access for this network? If so how do you plan to get the Internet routing and address translation that it will need?
HTH
Rick
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide