Tuesday, April 2, 2013

High-Density Wi-Fi Design Part 3 - WLAN Configuration Best Practices

In this video, I explain the best practices for configuring a Wi-Fi network for high-density environments. These include:
  • Proper encryption required to use 802.11n high throughput data rates
  • Proper use of Quality of Service (QoS) through Wi-Fi Multimedia (WMM)
  • Disabling lower data rates to maintain high performance
  • Prioritizing key business applications over recreational applications
  • Client rate-limiting to prevent "greedy" clients from hogging bandwidth
  • The important role that bi-directional band steering plays in optimizing spectral use
  • Load balancing clients based on airtime utilization on different channels to serve users where the most capacity exists
  • Using airtime fairness to adequately handle a mixed-client environment
  • Proper consideration of wired network resources, including switch port bandwidth, power over Ethernet, and Internet/WAN bandwidth
  • Appropriately sizing IP subnets to account for device density and user mobility


These principles are covered in more depth in the Aerohive High-Density Wi-Fi Design and Configuration Guide.

Read the Entire High-Density Wi-Fi Design Series:
Design Your WLAN for High Capacity
Video Blog: High-Density Wi-Fi Design Part 1 - Forecasting AP Capacity
Video Blog: High-Density Wi-Fi Design Part 2 - RF Planning
Video Blog: High-Density Wi-Fi Design Part 3 - WLAN Configuration Best Practices

Cheers,
Andrew

6 comments:

  1. Have you run across a wireless-only office - all end user devices using WLAN as primary connectivity instead of Ethernet? What devices would you recommend stay wired (network printers, VOIP phones)? Would this not also require high density of APs?

    ReplyDelete
    Replies
    1. Hi Dale,
      Yes, I've actually designed and implemented networks where wireless access is used for the majority of connections. It's never really an "all-wireless" or "wireless-only" office because some shared devices still need to be wired, as you mentioned.

      The most common devices that I still see being wired include:
      - Printers, because enterprise multi-function printers are often in shared areas. And their embedded Wi-Fi capabilities are usually very poor and perform horribly which would cause support issues.
      - A few ports in conference rooms for portable video carts and other temporary or high-bandwidth uses. Most end-users are still wireless though for meetings.
      - Executive offices still get wired access (in addition to wireless) because they require a higher level of service than cubicle workers.
      - VoIP desk phones that are cheaper than wireless VoIP phones. Although many offices are trying to move to soft phones on laptops to get around the cost difference.
      - Security cameras

      Yes, a mostly wireless office typically requires a higher density of APs.

      Cheers,
      Andrew

      Delete
  2. You mention that the sizing of the broadcast domain is a consideration for designing high capacity networks. I've seen some very large events use multiple /20s for their subnet sizes, but I'm curious if you've seen any larger subnets used and if there was any negative impact?

    Thanks,

    Jason

    ReplyDelete
    Replies
    1. Hi Jason,
      I've typically seen /20 or /21 subnets used. I haven't personally seen anything larger, and I would hesitate to do so personally.

      Cheers,
      Andrew

      Delete
    2. I've seen some of those new "plug and play wireless AP" that use /8 subnet for wireless client.
      Each AP has its own dhcp server.
      Do you think this could cause trobles?

      Delete
  3. hi Andrew,
    i am implementing wireless service provider(offload) on the public area, and unfortunately, so many rogue AP that detect on this area. I test the performance and it's really bad. What should i do to attack this interference. i am using Cisco product. am i customizing the RRM profile? *Note : i use the AP without clean air feature.

    ReplyDelete