Recommended way for routing vlans and choosing appliances for 600 users

[George-Sl]

Hello

Let's say we have 600 users divided in 6 vlans(each 100 hosts), what appliances do we use in real world?

1) is it gonna be all catalyst l3 switches with 600 ports?

2) or L2 switches with 600 ports that has trunk port to one of l3 switch interfaces?

if #1 is true do we typically use 45xx or higher in real world or 35XX?

if #2 is true how can you remove the bottle neck? I think #2 is not practical even with Link aggregation??

Thanks

[Leo Laohoo]

---

@George-Sl wrote:

if #2 is true how can you remove the bottle neck? I think #2 is not practical even with Link aggregation??

---

What bottleneck? 

With the advent of 1- and 10 Gbps uplinks, the only bottleneck I is ... actually, what bottleneck? 

What kind of office user(s) can actually have a sustained use of >600 Mbps?

[Dennis Mink]

Use cheap layer 2 switches on your access layer and let your core switches do the inter vlan routing. Use nexus switches at your core with vpc portchannels to your access switches.

[balaji.bandi]

4500 and 3500 almost end of life.

it all depends on budget and what feature you looking.

if you looking chasis and redudanty control plance, i would looking chasis model.

if not i would also consider stacking with lower model.

again depends on where this switch will be install in the path, core or distribute or access.

look at new models of Catalyst 9K. for long life and ROI.

[Joseph W. Doherty]

"Let's say we have 600 users divided in 6 vlans(each 100 hosts), what appliances do we use in real world?"

Your answer depends on what those 600 users are doing across the network.

BTW, #2 is an unusual topology having all the downstream ports feed into just a single L3 switch port. You mention link aggregation, but again it would be unlikely to have just one LAG logical link feeding all the downstream devices. More likely is some form of star topology off the core device which takes advantage of more ports than even a LAG supports.

As to what's used in the "real world", that depends on, again, what your users are doing across the network, and more often than not, what's the budget you've to work with.