In this section, Lisa Phifer explains how your 802.11n migration provides support for legacy client devices that are already used for best-effort data applications like Internet access and email, identifying best practices for migration. After reading this section, you will understand how to avoid disruption during an 802.11n migration, while still improving performance for newer 802.11n devices.
Don't miss any of the articles in this tutorial on deploying 802.11n for improved wireless LAN performance:
- 802.11n for improved WLAN performance
- 802.11n WLAN design best practices
- Supporting legacy devices
- Optimizing 802.11n for voice over WLAN
- Delivering video over your 802.11n WLAN
802.11n specifies protection mechanisms to coexist with 802.11a/g wireless LANs (WLAN). If you are already using 11a/g for best-effort data applications like Internet access and email, enable these mechanisms to avoid disruption. However, there are a few tricks to doing so while letting newer clients and more demanding applications make the most of your 802.11n migration.
- Modes of operation: In greenfield WLANs, 11n can use a more efficient high-throughput (HT) modulation understood only by other 11n devices. But when old and new devices contend for the same channel, everyone must know when it is busy to avoid collisions. To enable this, all 802.11n access points (APs) also support HT-Mixed-Mode, sending a non-HT preamble followed by a new HT preamble. This mode should be used on every channel shared by legacy devices (yours or neighbors’), but it can be disabled on channels used only by 11n devices.
- 20/40 coexistence: To be good neighbors, 11n devices must sense whether both halves of a bonded channel are free before transmitting. 20/40-capable clients can transmit 40 MHz frames over the bonded channel or 20 MHz frames over the primary half of a bonded channel. 20/40-capable APs that detect other APs using the secondary half of a bonded channel must revert to 20 MHz operation or tune to a different bonded channel. This disruption may be considered a necessary evil but can often be avoided by using channel bonding only at 5 GHz.
- Phased coexistence operation (PCO) mode: This option lets an 802.11n access point simultaneously support old and new clients by alternating between 20 and 40 MHz channel operation. But PCO adds a lot of overhead and delay, causes jitter in latency-sensitive applications, and is thus rarely implemented or recommended.
- Dual CTS protection: To avoid collisions with legacy 11b devices, 11g devices had to precede every data frame with RTS/CTS or CTS-to-Self control frames. 11n extends this with dual CTS protection, sending control frames in both HT and non-HT formats. This overhead may be necessary in some cases but can often be avoided by not sharing channels with legacy clients.
Clearly, band and channel planning play important roles in coexistence and migration. For example, consider leaving legacy data applications on existing Service Set Identifiers (SSIDs) in the 2.4 GHz band while deploying new mobility applications on SSIDs in the 5 GHz band. However, even dual-band clients can connect to only one AP radio/channel at any time, so plan new 5 GHz SSIDs to address all application needs.
To facilitate this, many 11n APs support band steering – proprietary techniques used to encourage dual-band clients to connect at 5 GHz when an AP advertises the same SSID on both bands. Band steering can transparently encourage client migration, but it can also increase roam time and reduce data rates for distant clients with borderline signal.
This was first published in December 2010