You did all that work to get your controller configured, tuned, and finally is stable. Congrats. Problem is, I just tripped over the power cord. Now what happens?
The answer depends on what mode the AP is in, as you know. With a “Local” mode AP deployment, if the controller goes away, so does connectivity. So in that scenario you want to make sure there’s some sort of HA deployment. With our “HA” part numbers, you pay basically for the hardware and no licenses. With this model you’ve got 2 options. N+1 or 1+1. The 3rd option is FlexConnect.
With N+1 each AP has a primary, secondary, and tertiary controller. If the primary goes down, it will join to the secondary, then tertiary. When the joined controller goes down, you’re out of luck for about 2-3 minutes. You either need fully licensed controllers or, as long as the AP’s don’t use it as primary, the HA part number can act as secondary or tertiary.
With 1+1, also know as Client Stateful Switch Over (Client-SSO), you have 2 controllers that are physically together and connected with a TP cable. One is the HA part number. The 2 controllers will constantly share Client-level state information and if the primary goes down, the secondary takes over in less than 300 mS.
When AP’s are in FlexConnect mode, the AP is responsible for the data-plane and therefore if the controller goes away, even for an extended period, will still function. Of course there’s some feature loss when you use FlexConnect such as mDNS. FlexConnect is often used in branch deployments, which typically means mDNS wouldn’t be necessary anyway.
Here’s a link to the Configuration Guide describing N+1 and 1+1 High Availability: http://www.cisco.com/c/en/us/td/docs/wireless/controller/8-0/configuration-guide/b_cg80/b_cg80_chapter_01110000.html
Here’s the description of FlexConnect: http://www.cisco.com/c/en/us/td/docs/wireless/controller/8-0/configuration-guide/b_cg80/b_cg80_chapter_010001000.html