When Good Wireless Feels Bad

If my client device doesn’t connect to your WLAN, your network must have a problem.

My iPad keeps getting dropped by your network.

I keep losing my Internet, your network sucks.

Ever hear anything along these lines? Sure, sometimes wireless networks do have problems. Access points crap out. Controllers fail. A switch glitches, and PoE isn’t sent to an AP. But on enterprise-grade hardware running proper code, these sorts of issues should be the exception. At the same time, even when “the problem” lives on the client device itself, it still feels like a network issue to the user.

With a daily load on my own WLAN that peaks around 16K, I see every kind of client device under the sun. Thankfully, we have a generally very healthy environment despite the relative complexity that comes with supporting any and every device type in a multi-SSID/security type environment. But trouble does hit the individual user on occasion; hence the purpose of this blog.

Even when the WLAN is running perfectly at each cell and all the way through the network’s important parts (DHCP, DNS, RADIUS, credential store, routing, etc), these are among the many factors can still make the wireless network “feel” crappy to individual clients:

  • OS upgrade causes trouble in wireless adapter
  • Wireless driver dated, needs update
  • Windows wireless driver not best fit for client, need Intel/Broadcom latest version
  • IPv6 getting in the way of IPv4
  • Client “sticks” to APs that common sense says it shouldn’t
  • On dual-mode devices (cellular data and Wi-Fi), each side of device occasionally causes trouble for the other
  • Client device requires legacy data rates not supported by WLAN
  • Client supplicant for 802.1x network gets corrupted, mis-configured
  • Local interference (usually in 2.4 GHz) causes issues
  • Client device clings to weak/poor 5 GHz connection when solid 2.4 GHz available
  • Client device has static IP address set from previous network use
  • User changes network password but doesn’t update supplicant config
  • Too small of an Internet pipe for user load
  • Trouble on the Internet, out in ISP land, impacting specific destinations
  • Client device is laden with malware that gets in the way of Internet access

You get the picture… there are many conditions that can impact the individual client, or a specific group of like client devices, and what worked yesterday may have been changed today by an OS update or patch.

Thankfully, when critical network building blocks do fail, we can either rely on our good instrumentation (you have that, right?) to tell us we lost a switch, or controller, or AP, etc. Or we can correlate based on good trouble report gathering (always happens, yes?) that there is something similar among users having issues- maybe a common AD grouping that RADIUS services are  borking on or the like. Good logs help, too.

Regardless of what is causing the pain, many clients instantly blame the network. Some can’t fathom that their shiny, expensive device could be imperfect in any way or that the mothership would ever send them a patch that wasn’t properly QA’d. It can be frustrating, but is also just part of the wireless support experience.

Things get easier if you have the rare environment where client types are tightly controlled and the BYOD water has yet to spill over the dam. For the rest of us, being aware of not only the health of the network but also of the various ills that can hit the client end of wireless (and what to do and how to communicate about them) is an absolute must. 

At the recent Wireless Field Day 4, I discussed this topic with my fellow delegates in a conference room in Building 4 of the Cisco Campus in San Jose.

Here’s a bit more on specific frustrations with the WLAN, from factors that are largely out of the admin’s hands.

6 thoughts on “When Good Wireless Feels Bad

  1. Keith R. Parsons

    Thanks Lee for leading that discussion. Those of us on the enterprise side get blamed for the ‘Network is Down’ way too often. Look to the client first would be a good rule of thumb. (right after ‘Check the Physical’.

    Keith

    Reply
  2. Pingback: When Good Wireless Feels Bad

  3. Pingback: Twitter is Killing The Helpdesk | wirednot

  4. Veli-Pekka Ketonen

    Nicely summarized, Lee. Not everyone sees the big picture, when you multiply the number of different configuration combinations in live network you get a BIG number.

    Veli-Pekka

    Reply
  5. Pingback: Loose Lips Sink Ships (and Network Credibility) | wirednot

Tell me what YOU think.

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s