As we know there are three layer in cisco Network Model:
So my question is in Core / Distribution layer should i use Etherchannel between switches or use Stacking if switches are stackble.
For ex: suppose I have two cisco 3750 switches . so should i use etherchannel between them or use stacking in core layer?
What are the advantages and disadv of both.
Are you using Cisco 3750 Switch at CORE layer also?
If yes I propose to use stacking between two switch.
With cascaded switches, each device is configured independently. With stacked switches, one is in charge of the stack and that's the one where all configurations are performed. So if you take two 48-port switches and stack them, you'll appear to have one giant 96-port switch when it comes to the configuration.
When we look at stacking, like I said, the difference is that this connection is part of the backplane instead of the spanning tree. So you're supercharging your switch setup and making it appear as a larger switch with more ports, but all under a single administrative control.
This is by far the simpler method of doing things, because configuration is centralized, and your biggest problem may be figuring out which switch is the controlling device! The upside is that it's easy and you don't lose ports to connect things together. The downside is that if one of the switches in the middle of your stack fails, what happens? Different switches and vendors actually treat this differently.
With Cisco StackWise (3750), if the master switch fails, another will simply take over. If by chance you get a split in your stack from failures, then you'll simply end up with two mini-stacks with two devices in charge. It's all part of Cisco's high-availability programs moving out to the wiring closet.
You do not need to bother about spanning-tree between two stackble switches.
When you connected single non stackble access swtich with the stack than you can use etherchannel and eliminate the use of spanning-tree.
Hope this will help to take decesion.
Rate of it will be helpful.
Depending upon how your final topology looks like, you can use either of these options. How does your final topology look like. Do you want to use the 3 tier architecture or a collapsed core with core/distribution on a single box.
For the first part, I would use a stack at the distribution layer to eliminate STP and run L3 to the core. All the link will be active and forwarding from the l2 and L3 perspective.
For the second part, you can be fine either the stack option or ether channel option running HSRP.i would use stacking as it gives me a single management plane and also no more STP. If you want to use HSRP topology, you have to manually tweak the STP/ vlan config so that you can use active/active topology.else STP will block half of the avian label bandwidth.
Hope this helps.
Sent from Cisco Technical Support iPad App
It all comes down to how you design the network.
In some design a rock solid no downtime network is reqired, Then you can not stack some solutions and some solutions you must stack.
Other times you might need speed but you have ample time during scheduled outages and stacking becomes a very good choise.
So stack or no stack is all dependant on your design and your requirements.
as a general rule if you stack you need to make sure you have ample time to change software and so on, remember a 3750 cant take some time to upgrade.