Manage Radio QoS Policies

Without a dedicated QoS policy, any wireless network operates on a best-effort delivery basis, meaning all traffic has equal priority and equal chance of being delivered in a timely manner. When congestion occurs, all traffic has an equal chance of being dropped!

When configuring a QoS policy for a radio, select specific network traffic, prioritize it, and use congestion-management and congestion-avoidance techniques to provide deployment customizations best suited to each QoS policy's intended wireless client base.

WiNG managed controllers and their associated access point radios and wireless clients support several Quality of Service (QoS) techniques enabling real-time applications (such as voice and video) to coexist with lower priority background applications (such as web, email, and file transfers). A well designed QoS policy should:

Classify and mark data traffic to accurately prioritize and segregate it (by access category) throughout the network.
Minimize the network delay and jitter for latency sensitive traffic.
Ensure higher priority traffic has a better likelihood of delivery in the event of network congestion.
Prevent the ineffective utilization of access points degrading session quality by configuring admission control mechanisms within each radio QoS policy.

In a wireless network, wireless clients supporting low and high priority traffic contend with one another for access and data resources. The IEEE 802.11e amendment has defined Enhanced Distributed Channel Access (EDCA) mechanisms stating high priority traffic can access the network sooner then lower priority traffic. The EDCA defines four traffic classes (or access categories): voice (highest), video (next highest), best effort, and background (lowest). The EDCA has defined a time interval for each traffic class, known as the Transmit Opportunity (TXOP). The TXOP prevents traffic of a higher priority from completely dominating the wireless medium, thus ensuring lower priority traffic is still supported by the controller or service platform, their associated access points and connected radios.

IEEE 802.11e includes an advanced power saving technique called Unscheduled Automatic Power Save Delivery (U-APSD) that provides a mechanism for wireless clients to retrieve packets buffered by an access point. U-APSD reduces the amount of signaling frames sent from a client to retrieve buffered data from an access point. U-APSD also allows access points to deliver buffered data frames as bursts, without backing-off between data frames. These improvements are useful for voice clients, as they provide improved battery life and call quality.

The Wi-Fi alliance has created Wireless Multimedia (WMM) and WMM Power Save(WMM-PS) certification programs to ensure interoperability between 802.11e WLAN infrastructure implementations and wireless clients. An access point managed wireless network supports both WMM and WMM-Power Save techniques. WMM and WMM-PS (U-APSD) are enabled by default in each WLAN profile.

Enabling WMM support on a WLAN just advertises the WLAN's WMM capability and radio configuration to wireless clients. The wireless clients must be also able to support WMM and use the values correctly while accessing the WLAN.

WMM includes advanced parameters (CWMin, CWMax, AIFSN and TXOP) specifying back-off duration and inter-frame spacing when accessing the network. These parameters apply to both connected access point radios and their wireless clients. Parameters that affect access point transmissions to their clients are controlled using per radio WMM settings, while parameters used by wireless clients are controlled by a WLAN's WMM settings.

Access points support static QoS mechanisms per WLAN to provide prioritization of WLAN traffic when legacy (non WMM) clients are deployed. An access point allows flexible WLAN mapping with a static WMM access control value. When enabled on a WLAN, traffic forwarded from to a client is prioritized and forwarded based on the WLAN‘s WMM access control setting.

Note

Statically setting a WLAN WMM access category value prioritizes traffic to the client, but does not prioritize traffic from the client.

Wireless network administrators can also assign weights to each WLAN in relation to user priority levels. The lower the weight, the lower the priority. Use a weighted round robin technique to achieve different QoS levels across WLANs.

You can rate-limit bandwidth for WLAN sessions. This form of per-user rate limiting enables administrators to define uplink and downlink bandwidth limits for users and clients. This sets the level of traffic a user or client can forward and receive over the WLAN. If the user or client exceeds the limit, excessive traffic is dropped.

Rate limits can be applied externally from a RADIUS server using Vendor Specific Attributes (VSAs). Rate limits can be applied to users authenticating using 802.1X, captive portal authentication and devices using MAC authentication.