Distribution Layer Necessary?

[jaredo_1]

We currently have 2 2848's operating as our core with 6-8 access switches and no vlans. I'm planning on adding a new core using 2 3500yl's with Adv License for inter-vlan routing and pushing the existing 2848's as a distribution layer so that I can use 24 port 3500yls' vs. 48 port to save costs.

My question is will there be any performance degradation or complications in the configuration with that type of setup, assuming VRRP at the core and MSTP/Vlanning?

Attached a diagram of the proposed setup.

[Patrick Terlisten · Accepted solution]

Hello,

you should provice enough links/ trunks from distribution to core. You should route at the distribution level, not core. If you route at the core, the whole traffic has to pass the distribution level. In this case, you can drop this level and attach the edge switches directly to the core.

Regards,
Patrick

[jaredo_1]

The 2800s at the distribution layer can't do inter-vlan routing/vrrp, and I don't have enough ports on the 3500s to get rid of the distribution layer. This setup should still work, I just want to make sure there won't be performance issues.

[jaredo_1]

This will also be a mstp setup with different vlan priorities on different distribution switches to load balance and separate traffic.

[Patrick Terlisten]

Hello,

in this case you should make sure, that the traffic is equal distributed over all uplinks to the core.

Regards,
Patrick

[jaredo_1]

Here is a link to a diagram containing the mstp information:

http://farm3.static.flickr.com/2709/4191026348_d4550742a6_o.jpg

[Patrick Terlisten]

Hi,

nice setup. This should work.

Regards,
Patrick

[Arran Cudbard-Bell]

Just a quick note: you shouldn't need to statically define instance priorities on the distribution layer, just the core (seemed like you were in the diagram).

[jaredo_1]

How come?

[jaredo_1]

I just assumed it would be necessary to define up to the distribution layer (not the access layer), otherwise wouldn't the access layer switches all route to one distribution switch instead of balancing between them?

[jaredo_1]

Looking at the core1 config (core2 mirrors this with slight differences in priorities and vrrp states):

hostname "CORE1"
module 1 type J86xxA
trunk 1-4 Trk1 Trunk
trunk 5-8 Trk2 Trunk
trunk 9-12 Trk3 Trunk
ip routing
vlan 1
name "DEFAULT_VLAN"
untagged 13-18,21-24
ip address 10.10.0.1 255.255.255.0
tagged Trk1-Trk3
no untagged 19-20
exit
vlan 10
name "Servers"
untagged 20
ip address 10.10.10.1 255.255.255.0
tagged Trk1-Trk3
exit
vlan 30
name "Printers"
ip address 10.10.30.2 255.255.255.0
tagged Trk1-Trk3
exit
vlan 100
name "DownstairsOB"
untagged 19
ip address 10.10.100.2 255.255.255.0
tagged Trk1-Trk3
exit
vlan 101
name "UpstairsOB"
ip address 10.10.101.2 255.255.255.0
tagged Trk1-Trk3
exit
vlan 200
name "Warehouse"
ip address 10.10.200.2 255.255.255.0
tagged Trk1-Trk3
exit
vlan 20
name "Reserved"
no ip address
exit
ip route 0.0.0.0 0.0.0.0 10.10.0.250
router rip
no auto-summary
exit
router vrrp
snmp-server community "public" unrestricted
spanning-tree
spanning-tree Trk1 priority 4
spanning-tree Trk2 priority 4
spanning-tree Trk3 priority 4
spanning-tree config-name "mstp"
spanning-tree config-revision 1
spanning-tree instance 1 vlan 1 10 20
spanning-tree instance 1 priority 0
spanning-tree instance 1 Trk1 priority 4
spanning-tree instance 1 Trk2 priority 4
spanning-tree instance 1 Trk3 priority 4
spanning-tree instance 2 vlan 30 100 101 200
spanning-tree instance 2 priority 1
spanning-tree instance 2 Trk1 priority 4
spanning-tree instance 2 Trk2 priority 4
spanning-tree instance 2 Trk3 priority 4
spanning-tree priority 0
vlan 1
vrrp vrid 1
owner
virtual-ip-address 10.10.0.1 255.255.255.0
priority 255
enable
exit
exit
vlan 10
ip rip 10.10.10.1
vrrp vrid 1
owner
virtual-ip-address 10.10.10.1 255.255.255.0
priority 255
enable
exit
exit
vlan 30
vrrp vrid 1
backup
virtual-ip-address 10.10.30.1 255.255.255.0
enable
exit
exit
vlan 100
vrrp vrid 1
backup
virtual-ip-address 10.10.100.1 255.255.255.0
enable
exit
exit
vlan 101
vrrp vrid 1
backup
virtual-ip-address 10.10.101.1 255.255.255.0
enable
exit
exit
vlan 200
vrrp vrid 1
backup
virtual-ip-address 10.10.200.1 255.255.255.0
enable
exit
exit



I noticed each spanning-tree instance assigns individual priorities to the trunks as well, does the priorities on the instance 1 trunks need to be adjusted at all so that the priorities are higher (less) than instance 0? Or does the overall higher priority of instance 1 really all that matters?

[jaredo_1]

Sorry, mean instance 1 & 2 not 0 & 1.