I apologize if this has appeared in another thread, but it has been something I've been mulling over now for a couple weeks. With the improvements in the 7.2.103, is there any reason to not automatically make every deployment an H-REAP (now FlexConnect) deployment? Just the mere fact of WLAN survivability during controller outage, without a redundant controller of course, is a strong incentive. With all the progress Cisco has made, I'm trying to come up with reasons to not just configure every deployment to be H-REAP/Flexconnect. Off the top of my head the only reasons I can think of are:
-extra config on the access switch
-limited to 25 AP's per H-REAP group. (Even this reason is really not that big of one when designed properly)
I think other would consider it more if there were no limitation when comparing h-reap to local mode. When there are redundant WLC's I wouldn't even think about using h-reap in my installs. Maybe if my customer only had one WLC and required some sort of uptime if the wlc fails. The extra commands is no big deal it just the "limitations" when using h-reap. I probably wouldn't even use h-reap groups if the AP's, radius and wlc were in the same location.
At this time I don't think the question should be why or why not use FlexConnect of every single deployment. Local mode and FlexConnect both have their own benefits and each should be carefully considered for the deployment in question. A few items to keep in mind are:
Where will the APs be located in relation to the WLC?
Who is going to be managing this system? (FlexConnect is more difficult to manage than local due to the additional configuration.)
Does it make sense for data traffic? (In one case I had a customer that did not have their controller located at a central point of their network... Not my choice or reccomendation.)
Do I need AAA override?
FlexConnect isn't terrible but it's still got a ways to go regarding development. It was designed around the idea of remote branch deployment, not necessairly large depolyments at one site where a customer could have multiple redundant controllers.
So far some limitations I've personally faced are bugs.
Don't let that stop you from utilizing FlexConnect as it does indeed have its uses.
For some good information check out the follwoing link:
I have a mixed deployment. I use HREAP where it makes the most sense...on data and voice WLANs at remote branch locations. Why should local data destined for a printer at a remote location traverse a WAN cloud? Why shouldn't wireless VOIP devices be on the same voice VLAN as everyone else at the remote branch location?
I also have WLANs that are centrally switched. The traffic is destined back to the controllers at the headquarters location anyways....for instance, anchored WLANs that are carried for other government agencies, or an internet only WLAN for guests.
I use AP groups to control which APs carry which WLANs.
I have this setup on redundant 5508 controllers (in the same location). I have tested it and it works very well. For reduncancy, your WLANS must have the same numbers and you must have the same AP groups available on both controllers.