This vendor-written tech primer has been edited by Network World to eliminate product promotion, but readers should note it will likely favor the submitter's approach.
The latest challenge confronting IT organizations is managing the influx of mobile devices and real-time applications that rely on the enterprise wireless LAN. The WLAN has to be resilient enough to enable mobile devices to stay connected and applications like voice and videoconferencing to work without interruption.
To increase resiliency in any network infrastructure, it's good to start by employing a network design ready for high availability. But designing for high availability is not always enough to ensure WLANs are ready for mission-critical applications.
ANALYSIS: It's time for wireless LANs to evolve
In wired access networks, we hardly worry about the user devices disconnecting from the network in case of a network failure, because the cable always stays connected as recovery takes place within the redundancy infrastructure. With WLANs this can be a challenge because traditional wireless access points (APs) turn off their Wi-Fi radios during a controller failover and mobile devices then start to search for a new AP.
So the easiest way to achieve a wired-like resiliency for WLANs is to not give mobile devices any chance to roam -- that is, to never drop them off the Wi-Fi network until WLAN recovery is complete. To achieve this, wireless APs should continue to broadcast SSIDs so mobile devices won't resort to scanning for other APs or SSIDs to connect to. This greatly reduces the possibility that a mobile device will disconnect from the network.
Without this approach end users are forced to manually adjust (i.e., physically turn Wi-Fi off and back on again) their mobile devices to re-establish connectivity, resulting in unsatisfied Wi-Fi users and potentially increasing IT support costs. The easy test when you are evaluating WLAN gear -- the WLAN infrastructure should not disconnect any mobile devices after a redundancy failover event takes place.
A high-availability approach that does not drop devices will also deliver sustained connectivity for real-time mobile applications as well. Video streaming, unified communications and voice/video calling are all examples of applications where end users expect consistently reliable service and don't want to be asked to re-initiate a voice/video call. We've all experienced this frustration with dropped connections on cellular networks. As recovery times for WLANs decrease, real-time mobile applications have a greater chance to stay connected in case of network failures without requiring manual intervention by the end users.
Therefore, another valuable test would be to measure amount of time (and effort) it takes to recover any real-time application after a simulated failure within the redundant WLAN infrastructure.
It is important to mention that this type of resiliency should not come at an increased cost. Most WLANs support traditional N+1 redundancy where one controller offers protection for many others. This type of high-availability design reduces maintenance expenses, and provides significant cost savings in large-scale deployments.
Delivering consistent performance for high-density
Traditionally, as the number of mobile devices on a wireless network increases, total available bandwidth per AP tends to decline. Again, this is not a concern in wired access networking; we hardly ever worry about total performance of an access switch decreasing as we connect more PCs to it. Given Wi-Fi is fast becoming the primary network access method, it makes sense to expect a similar level of consistency in performance from WLANs.
If a WLAN is not designed for high-density, the end user can experience what's usually referred to as a "Wi-Fi meltdown," an unexplained halt in network connectivity. Choice of antenna and AP types, proper placement of the APs and appropriate channel planning for increased capacity are all important design criteria, but proper RF design is not always enough to ensure predictable and acceptable performance for mobile devices.
In high-density environments, RF management techniques implemented within the WLAN infrastructure also play a big role. First, the APs need to ensure proper "airtime" allocation, ensuring bandwidth availability for all. For instance, an Apple MacBook Pro installed with a 3x3:3 MIMO Wi-Fi chipset (450Mbps max Wi-Fi link speed) should achieve higher performance compared to the newest Apple iPad A installed with a 1x1:1 MIMO Wi-Fi chipset (65Mbps max Wi-Fi link speed) when both are associated to the same AP. In other words, faster clients should get their job done faster so that there is more airtime available for everyone else. No single group of mobile devices should be able to monopolize network resources.
Wireless engineers can learn a lot about the behavior and expected performance of mobile devices within their infrastructure after simple throughput measurements. Testing the upload and download performance simultaneously is highly recommended since most mobile devices are now powered by cloud-based mobile applications which upload data to the network as much as they download.
Secondly, APs should take into account the mobile device behavior in high-density environments. End users, and hence mobile devices, hardly move in locations such as auditoriums, classrooms, company events and meetings. Ideally, their Wi-Fi link speeds should stay consistent when communicating with the APs.
Traditionally, APs employ algorithms that reduce the Wi-Fi transmit rate, using 243Mbps instead of 300Mbps, for example, to individual mobile devices in case of packet loss. This is done to increase reliability of the Wi-Fi link with the assumption that the mobile device is on the move -- in other words, moving away from the AP. However, in a high-density environment, the packet loss might result from contention given the number of mobile devices trying to access a specific Wi-Fi channel.
Hence the "rate vs. range" algorithms used to enable seamless mobility for mobile devices as they roam between APs may not necessarily be suitable for high-density Wi-Fi. In fact, if not taken into account properly, they may make things worse: The more APs reduce their link speed while talking to a mobile device, the more airtime they consume. This, in turn, will mean less airtime availability for all, creating the perfect recipe for a Wi-Fi meltdown.
At the end of the day, end users just want a Wi-Fi network that works. In healthcare, quality of service for mission-critical applications is a must -- no exceptions. In retail and warehousing, disruptions within the Wi-Fi network will mean substantial loss of revenue. In the general enterprise, most employees do not even have a way to connect to the wired network anymore as the access network moves to primarily wireless. Without high performance Wi-Fi, no work gets done.
To enable this seamless experience for end users and their mobile devices, wireless network engineers should demand more from their WLAN infrastructure. Consistent, predictable experience on the Wi-Fi network should be a table-stakes requirement.
Read more about lan and wan in Network World's LAN & WAN section.
This story, "Ensuring Resiliency in Wireless LANs" was originally published by Network World.