asked on

HP 3800 distributed trunking

Hi,

Hoping someone could shed some light on this subject as extensive reading doesn't give a clear answer.

I need to have resilience in case of switch failure and also use the bandwidth
I have 2 x HP 3800 switches. I have my core HP servers in the same rack. I plan on using 1 server NIC for each switch (using 2 in total) I will be bonding the server NICs using mode 4 (they are Linux servers)

Question: Do I need to use stacking modules or can I just run 1gb connection between 2 of the switch ports to 'join' the 2 switches?
Does using stacking ports make the resilience better?

Regards

Zephyr ICT

So it's basically stacking vs trunking (LAG), stacking should in theory be faster because it can use the complete speed of the backplane as well as give you benefit in ease of management (depending on the stacking technology) because you'll have only 1 configuration to maintain as the stacked switches will behave as 1.

I'd say resilience wise their is less of a difference, but I'm thinking still a benefit for stacking.

If I had the choice I'd go for stacking.

Now for all disclosure, this is a generalization, I haven't read up on this particular switch, but it's mostly the case concerning stacking

noci

You cannot build a Trunk spanning two independent switches,
So stacking is required if you want to use two different switches to connect one server to.

nik middleton

ASKER

According to HP's docs, Distributed trunking does exactly that. My question was, is stacking mandatory for best performance/reliability ? So much about stacking just seems to mention it's only 1 switch to manage and that's the main USP

ASKER CERTIFIED SOLUTION

Zephyr ICT

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial