Need Help to Setup Core Switch Redundancy Using HSRP

[Ron Timbang]

To All,

Good day!

Need help to setup redundancy for our core switch.  I have attached our network diagram. Basically we have one core switch & no redundancy in case of failure. Every server is also connected to the switch. Any suggestions on how to setup a redundant network.  Budget is also a concern. Please also suggest switch model we can use. We have 12 servers & 150 workstations. I have attached a proposed setup, planning to use HSRP. Please let me know if you have suggestions.

Do i need to add additional layer - > Distribution -> Access or it's ok to connect servers directly to the core.

PS. Can i setup the server with 2 fe connection. 1 -> primary & 2-> secondary with same ip on both nic? 

Thanks in advance..

[kapilgupta1]

Hi Ron,

it’s not recommended to have everything hooked onto one core switch. An ideal network should have redundancy built in. I suggest you to have a collapse backbone model wherein you can combine Core & Distribution functionality on core switches. I also suggest having separate server farm switches (may be C3750X-24 having 1 gig ports) to servers connected to Core switches. this will provide you redundancy for server as you can connected server network ports on alternate server farm switches and have cris-cross connectivity between server farm and core on OSPF. This will give advantage of limiting any issue within server farm layer and your remaining network will be unaffected.

I can send you a sample diagram on how your network can be look like but budget will be a constraint to build a good level redundancy. Let me know if you want me to send a model diagram or any further details.

Regards,

Kapil

[Ron Timbang]

Thank you Kapil, appreciate your help. Please send me the diagram. 

We have static routing configured on the core, in-case of HSRP do I need to configure the same static routing on the secondary switch? How about the Gigabit link? Is that enough or fiber connection is recommended for the connection to the core? or etherchannels? 

Again thanks a lot!

[kapilgupta1]

Hi Ron,

There are several ways to design your network and i can further help if you can share more information about your network devices models, their capabilities (L2, L3) etc. I am enclosing a sample topology diagram for your reference but again this is not the only way to design your network but yes this design will definately provide good level of resiliency in your network.

We can further discuss on this if are are confortable in sharing more information about your existing network and your expectations from proposed network.

Regards,

Kapil

[Joseph W. Doherty]

What's your core device now (in detail)?

[dilip.sharma2005]

No distribution layer required, Just core and access switch. Core switch with same config running HSRP between them. 

Access switch connected to both core switch.

You have to check your server support NC teaming or not if yes you can configure one single ip on a pair (virtual nic) and will act as single NIC. Physicaly both NIC will be connected to different core switch.

This will provide you what you want to achieve .

Dilip