Wireless Architectures and the WLC
One access point is easy. Two hundred of them, spread across a campus, all sharing one SSID and one security policy, are not. This final lesson of Week 4 shows how wireless is architected: from a lone self-managed access point, to a fleet of lightweight APs steered by a central Wireless LAN Controller, to management that lives in the cloud. It closes with the AP modes and deployment choices that make a wireless design actually work.
Why wireless architecture matters
A wireless network is not just radios; it is a system. Each access point must serve clients on the right SSID, apply the right security, hand them onto a VLAN, and cooperate with its neighbours so a moving client can roam without dropping its session. Doing that consistently is a management problem as much as a radio problem, and the way you solve it defines your wireless architecture.
There are two poles. In the autonomous model, every access point is a small, self-sufficient device you configure one by one. In the controller-based model, the access points are deliberately made “lightweight” — dumb about policy, smart only about radio — and a central Wireless LAN Controller (WLC) holds the configuration, security, and RF intelligence for the whole fleet. A third, newer option moves that central brain into the cloud. The trend across the industry is unmistakable: from standalone, to on-premises controller, to cloud-managed.
Autonomous access points
An autonomous access point (sometimes called standalone) runs the full wireless stack itself. It holds its own configuration, advertises its own SSID, authenticates its own clients, and bridges them to the wired LAN. You manage it by console, Telnet/SSH, or a web interface, and you repeat that work for every single AP.
Autonomous APs are simple and robust — there is no controller to fail and no tunnel to maintain — which makes them fine for a home, a small office, or a branch with one or two APs. They scale badly: changing the Wi-Fi password means touching every device, there is no central view of the RF environment, and roaming and security consistency are hard to guarantee. These are exactly the pains the controller-based model was invented to remove.
No controller to fail, no CAPWAP tunnel to maintain, no licence or platform to buy. The AP keeps working on its own if the management network goes down.
Every change is repeated on every device. No central view of the RF environment, inconsistent roaming, and security drift as one AP is missed during an update.
Controller-based WLANs and the WLC
In a controller-based deployment the access points are lightweight (Cisco calls them LWAPPs, or lightweight access points). They no longer hold the network configuration. Instead they join a WLC, which becomes the single point of configuration, monitoring, security policy, and RF management. Configure one WLAN profile on the WLC and every joined AP offers it.
The WLC also presents several logical interfaces you should know by name: the management interface (for administering the controller, usually on a management VLAN), the dynamic interfaces (one per client VLAN/WLAN), the AP-manager interface (handling AP joins), the virtual interface (used for mobility and DHCP relay), and on some models a service-port for out-of-band management. A wireless client’s traffic typically maps from an SSID to a WLAN on the WLC and then to a dynamic interface — and therefore to a VLAN.
Cloud-managed wireless is the same idea with the controller relocated: the management and policy engine lives in a vendor cloud portal rather than in a box on site. Between the two sits Cisco’s own management platform, Catalyst Center (formerly DNA Center), which can configure and monitor both wired and wireless as a single fabric.
Split-MAC: which job lives where
To make a lightweight AP useful, the 802.11 MAC functions are split between the AP and the WLC — hence the name split-MAC. The guiding rule is simple: real-time radio work stays at the edge; management and policy work moves to the controller.
| Function | Lightweight AP | WLC |
|---|---|---|
| Beacons and probe responses | Yes — time-critical | No |
| ACKs, retransmissions, frame timing | Yes — real-time 802.11 | No |
| Authentication, association, roaming | No | Yes — central policy |
| Security policy and key management | No | Yes |
| RF management, transmit power, channels | Assists | Yes — RRM decides |
| Client traffic to the wired LAN | Tunnels it (local mode) | Bridges it to the correct VLAN |
Lightweight APs carry no network configuration. They discover and join the WLC, then tunnel both their control traffic and (in local mode) their client data through it using CAPWAP over the wired LAN. The WLC applies authentication, roaming, security, and RF policy for every AP at once.
CAPWAP: discovery and join
The protocol that carries a lightweight AP’s traffic to the WLC is CAPWAP — Control And Provisioning of Wireless Access Points (RFC 5415/5416). It is the modern successor to the older Cisco LWAPP, and its ports are exam facts:
| CAPWAP tunnel | Transport | Purpose |
|---|---|---|
| Control | UDP 5246 | AP join, configuration, and management messages; encrypted with DTLS. |
| Data | UDP 5247 | Client traffic tunnelled to the WLC (optional DTLS encryption). |
Before any of this, the AP must find the controller. A booting AP gets an IP address from DHCP, then performs CAPWAP discovery using, in order of preference:
- DHCP option 43 — the DHCP server can hand the AP the WLC’s IP address directly. This is the usual method on an enterprise network.
- DNS — the AP can resolve a well-known name such as
CISCO-CAPWAP-CONTROLLER.<domain>to a controller address. - Broadcast / stored address — a local Layer 2 broadcast, or a controller address remembered from a previous join.
Once it finds the WLC, the AP and controller exchange discovery and join messages, establish DTLS security on the control channel, and the WLC downloads configuration. The AP is then ready to serve clients.
Cloud-managed wireless
Cloud-managed wireless (for example Cisco Meraki) removes the on-premises controller entirely. Access points connect out to a vendor cloud dashboard where the administrator configures SSIDs, security, and RF settings, and where statistics and monitoring are aggregated. There is no local controller appliance to buy, license, or maintain, and the network is manageable from anywhere.
The trade-off is dependence on the cloud connection and on the vendor’s platform. For many distributed organisations — many small sites, few IT staff — that trade is worth it. The exam level insight is simply to recognise the three models side by side: autonomous (per-AP, no controller), controller-based (lightweight APs plus an on-premises WLC), and cloud-managed (lightweight APs plus a cloud dashboard).
Access point modes
A Cisco lightweight AP can be placed in one of several modes, each giving the radio a different job. Know all six:
| Mode | What the AP does |
|---|---|
| Local | The default mode. Serves clients on its WLANs and tunnels traffic to the WLC. While not busy it scans other channels for rogue APs. |
| FlexConnect | For branch offices. The AP can switch client traffic locally at the branch while staying controlled by a central WLC, and can keep serving clients if the WAN link to the controller fails. |
| Monitor | Dedicated to listening. Serves no clients; scans all channels for rogue access points and wireless attacks (IDS/IPS). |
| Sniffer | Captures 802.11 frames on a chosen channel and forwards them to a packet analyser such as Wireshark, for deep protocol troubleshooting. |
| Bridge / Mesh | Provides wireless links between access points (mesh), including point-to-point bridging where cabling is impractical. |
| SE-connect | Spectrum Expert Connect: passes raw RF spectrum data to a spectrum-analysis tool to find interference and non-Wi-Fi sources. |
Deployment considerations and the bands recap
Deciding where and how to deploy APs is as important as choosing the architecture.
- Coverage versus capacity. A warehouse needs coverage — a few APs at high power to reach every corner. A lecture hall or open office needs capacity — many APs at low power so each serves a smaller “cell” and clients share less air. Dense spaces are a capacity problem, not a coverage problem.
- Site survey. Before installing, a survey (predictive or on-site) maps signal strength and interference so APs land where users actually are. Guessing the layout produces dead spots and co-channel interference.
- Channel planning. Neighbouring APs must use different channels; in 2.4 GHz that means the non-overlapping 1, 6, and 11 reused in a pattern. The 5 and 6 GHz bands have many more channels, so reuse is far easier.
- Transmit power. Lower power in dense areas shrinks cells, reduces co-channel interference, and encourages clients to roam at the right time. Too much power creates overlap and sticky clients.
- Roaming. All APs share one SSID so the network looks like one. A Layer 2 roam keeps the client on the same subnet and the same IP address; a Layer 3 roam moves it across subnets and requires mobility support on the WLC to preserve the session (important for voice).
- Guest and security design. Keep guest traffic isolated on its own WLAN and VLAN, use WPA2 (AES-CCMP) or the stronger WPA3 (SAE), and use 802.1X/EAP for enterprise authentication.
Bands and channels recap
Bring the week full circle. The radio bands decide range, throughput, and how much channel reuse you get:
| Band | Character | Channel reuse |
|---|---|---|
| 2.4 GHz | Best range and wall penetration; most interference | Only 1, 6, 11 do not overlap |
| 5 GHz | More throughput, less interference, shorter range | Many non-overlapping channels |
| 6 GHz | Cleanest, shortest range; enables Wi-Fi 6E | The most channels of all |
The design lesson is that architecture and radio are one design. A controller gives you consistent channels and power across the fleet through its RF management, but the physics — range against throughput, and the 1/6/11 constraint in 2.4 GHz — still sets the floor.
Review
Check what you have learned with these review questions.
Sources & references
- Calhoun, P. et al. RFC 5415: Control And Provisioning of Wireless Access Points (CAPWAP) Protocol Specification (2009). rfc-editor.org/rfc/rfc5415.
- Calhoun, P. et al. RFC 5416: CAPWAP Protocol Binding for IEEE 802.11 (2009). rfc-editor.org/rfc/rfc5416.
- Cisco. Wireless LAN Controller Configuration Guide (AP modes, CAPWAP, split-MAC, WLC interfaces). cisco.com.
- Cisco. CCNA (200-301) Exam Topics, Network Fundamentals — wireless architectures and AP modes. cisco.com.
Target exams: CCNA 200-301 v1.1 · CyberOps Associate 200-201. Last content verification: 2026-10-11.