1

I have been studying Flex Connect and Cisco WLC deployment and I noticed that the default mode for WLC's and LAP's (Light-weight Access Points) is centrally switched (not using Flex-Connect) which got me thinking. If every single frame from a wireless client is destined to somewhere on the LAN, the frame has to first go to the WLC even if the destination of the frame is literally connected to the same LAP, and if the WLC is located multiple-layer two hops away as it usually would be in a large enterprise setup, or even at a remote branch over a WAN, would that not increase the delay by a ridiculously high amount?

I guess what I am really asking is, if that's true, why would anyone NOT use Flex-Connect?

Reference material.

Thanks

2
  • "would that not increase the delay by a ridiculously high amount? Not really. It does increase the delay, but it works pretty well, and it gives the enterprise more control over the (possibly dangerous) traffic coming into the WAPs. You can have multiple WLCs that are in data centers closer to various branches. Cisco calls flexconnect a kludge. There are now various switches the have sub-WLCs built into them. and they connect to the main WLCs to be able to keep the traffic more local.
    – Ron Maupin
    Sep 27 at 12:30
  • So you’re aware, some of Cisco’s newer Aironet models have WLCs built into them, so you can have central management of all APs without a switch-based WLC. This comes in handy for environments with mixed brands (like HP switches with Cisco Aironets). The 2802i is one such model, for example.
    – Jesse P.
    Sep 27 at 20:07
1

Easier deployment, since you don't have to plan VLANs ahead and push them to switches. Management VLAN for APs in access mode is enough. Gives you the ability to use DNS ACLs.VLAN pooling needs to be done in centralized deployment. If you need inter controller mobility you can not have it in flex mode.

If you think about a normal enterprise customer with standard needs, you will see that most of the traffic will either flow through the internet or to the data center where the controller resides. There isn't much traffic between endpoints these days. on the contrary, we try to deny this traffic as much as possible. So the traffic is most likely to flow to the DC either offloaded or in CAPWAP.

All is said, most of my deployments are in flex mode. Because of the simple reason you stated. Why would I not use the full potential of the APs when I can and limit myself? Also, flex mode gives you flexibility in remote locations. But please note that even in flex mode, you may choose to implement tunnel mode in some WLANs. An example might be guest SSID (or Contractor SSID). Either in HQ or in the branch office, you may want to tunnel them to the controller and offload it to a more secure area where you can easily control the traffic.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.