ICX 7750 routing table confuses me

  • 1
  • Question
  • Updated 3 months ago
We have a 7750 as our core router and remotes sites with either a 7750 or 7450 as their router.  The configurations were set up by an outside vendor and they only answer about half my questions.

The core 7750 has routes something like:

ip route 10.10.0.0/16 ip_address_1
ip route 10.20.0.0/16 ip_address_2
ip route 10.30.0.0/16 ip_address_3
ip route 10.40.0.0/16 ip_address_4
ip route 10.41.0.0/16 ip_address_4

Each site has a wired network (10.10, 10.20, etc) and a wireless network (10.11, 10.21, etc).

My confusion comes from why some of the sites only have the wired route while others have both a wired route and a wireless route but everything works just fine.

As I understood it, having 10.10.0.0/16 as an ip route means that any packet with a 10.10.x.x address gets routed to ip_address_1.  So how do the 10.11.x.x addressed packets end up in the same place?  If I set up a 10.12.0.0 vlan, would it need it's own route?

Any explanation/clarification would be appreciated.
Photo of Clayton Tavernier

Clayton Tavernier

  • 65 Posts
  • 3 Reply Likes

Posted 3 months ago

  • 1
Photo of Roberto Hernandez Jr

Roberto Hernandez Jr, Employee

  • 19 Posts
  • 7 Reply Likes
Hi Clayton,

Your assumptions, based on the information you have provided, are correct. At lot of it depend on y our network architecture. If we take the routes that you show as the only routes in your table then, you are correct. Traffic to 10.11.10.21 would go no where. Unless there is a default route installed. It is possible, and not shown here, that there might be some dynamic route (OSPF, RIP). We would only be able to know that if you do a "show ip route".  Could you share that? It is also possible that policy based forwarding could be used. In this case, routing would not impact how packets move across the switch. In summary, we3 would have to look at your entire routing table and configuration to determine how things are being forwarded.
Photo of Clayton Tavernier

Clayton Tavernier

  • 65 Posts
  • 3 Reply Likes
Thanks for the reply.

There's roughly 300 lines for "show ip route" so I'll just share a piece but it does look like the routing is dynamic.

#show ip route
Total number of IP routes: 293
Type Codes - B:BGP D:Connected O:OSPF R:RIP S:Static; Cost - Dist/Metric
BGP  Codes - i:iBGP e:eBGP
OSPF Codes - i:Inter Area 1:External Type 1 2:External Type 2
        Destination        Gateway         Port          Cost          Type Uptime
1       0.0.0.0/0          10.206.1.1      e 1/1/5       1/1           S    87d1h 
2       10.10.0.0/16       10.255.254.54   e 1/1/20      1/1           S    41d20h
3       10.11.0.0/23       10.255.254.54   e 1/1/20      110/2         O    41d20h
4       10.11.2.0/23       10.255.254.54   e 1/1/20      110/2         O    41d20h
5       10.11.8.0/23       10.255.254.54   e 1/1/20      110/2         O    41d20h
6       10.11.10.0/23      10.255.254.54   e 1/1/20      110/2         O    41d20h
7       10.11.12.0/23      10.255.254.54   e 1/1/20      110/2         O    41d20h
8       10.11.14.0/23      10.255.254.54   e 1/1/20      110/2         O    41d20h
9       10.11.16.0/24      10.255.254.54   e 1/1/20      110/2         O    41d20h
10      10.11.17.0/24      10.255.254.54   e 1/1/20      110/2         O    41d20h
11      10.11.18.0/24      10.255.254.54   e 1/1/20      110/2         O    41d20h
12      10.11.19.0/24      10.255.254.54   e 1/1/20      110/2         O    41d20h
13      10.11.20.0/24      10.255.254.54   e 1/1/20      110/2         O    41d20h
14      10.11.32.0/20      10.255.254.54   e 1/1/20      110/2         O    41d20h
15      10.11.48.0/20      10.255.254.54   e 1/1/20      110/2         O    41d20h
16      10.11.80.0/20      10.255.254.54   e 1/1/20      110/2         O    41d20h
17      10.11.96.0/20      10.255.254.54   e 1/1/20      110/2         O    41d20h
18      10.11.112.0/24     10.255.254.54   e 1/1/20      110/2         O    41d20h
19      10.11.254.254/32   10.255.254.54   e 1/1/20      110/2         O    41d20h

Photo of Roberto Hernandez Jr

Roberto Hernandez Jr, Employee

  • 19 Posts
  • 7 Reply Likes
Ok. Yeah. They are using OSPF. Hopefully that answers your question


(Edited)
Photo of Clayton Tavernier

Clayton Tavernier

  • 65 Posts
  • 3 Reply Likes
Ok, I know next to nothing about OSPF but I can see why every route isn't defined in the switch configuration.  Ignorant question: if the routing is dynamic, why are any routes defined in the switch config?
Photo of Roberto Hernandez Jr

Roberto Hernandez Jr, Employee

  • 19 Posts
  • 7 Reply Likes
I would have to understand the entire architecture but sometimes you want specific routes to a static destination and you do not want it to be affected by the dynamic routing. There could be other reasons based on the implementation It is not unusual to have a mix. In theory, dynamic routing is preferred because its dynamic and its easier to maintain across the network. However, I cannot specifically say why in this case it is the way it is as it would require knowledge of the entire network architecture, and purpose to understand why it was designed this way.
Photo of Clayton Tavernier

Clayton Tavernier

  • 65 Posts
  • 3 Reply Likes
Thanks for taking the time to explain all this to me.

Last question: The section of the dynamic table that I shared is copied (other than the IP addresses) for most but not all of our sites.  So, knowing what little you know about our infrastructure, could you explain why the dynamic routes wouldn't be copied across all our sites?  In other words, why are some of the sites missing most of the dynamic routes if the vlans are defined on all our switches?
Photo of Roberto Hernandez Jr

Roberto Hernandez Jr, Employee

  • 19 Posts
  • 7 Reply Likes
That would be difficult to know. What I can tell you is that knowing that you are running OSPF, it should replicate across as OSPF (LSDB) database is replicated. However, there are feature in OSPF that may filter, or summarized routes, or not allow certain routes to be propagated. Also if the greater network (at the core) uses routing protocol, like BGP or ISIS then there could be further filtering of routes. My guess is that some of the sites may just have default routes since they may not have more paths than one (typically known as stubby areas in OSPF)
Photo of Clayton Tavernier

Clayton Tavernier

  • 65 Posts
  • 3 Reply Likes
Thanks again for taking the time.