How Real-World Stress Test Of High-Density Wi-Fi Deployment Worked
Last week we published an article providing the details of a live, high-density Wi-fi test conducted at a school campus. The goal was to simulate a standardized testing scenario in order to demonstrate that the WLAN can support a large group of students simultaneously taking online exams. This week we provide the details of that test.
Probably every one of you has been presented with “completely independent” multiple Wi-Fi vendor stress tests that claim to produce non-biased results. Upon closer examination, you find that the testing was actually constructed to produce favorable results for a specific vendor or product.
To that end, Aerohive decided to test enterprise-class features and validate recommended configurations for K-12 educational environments in a real school, with real student devices running tests performed by real students using real school applications.
Special thanks go to Chris Moghtaderi, District Technology Coordinator at the Healdsburg Unified School District for his time and patience with us, and to all the Healdsburg High School students that showed up in the middle of summer to participate for hours in the testing, for sandwiches and coffee, (but no ice cream), for their effort.
Environment : High density
Several different tests were performed at two different locations. Both locations, a cafeteria and a gymnasium, were chosen because these are challenging, high-density environments and because they are typical locations for large-scale online examinations
One challenge represented a high number of client devices that needed to be supported in a large open space, with no attenuators like classroom walls. The most important requirement was 100% network availability for client devices running critical test applications during the testing periods.
Evaluation criteria: Real life situations
Students were asked to connect to the student SSID and to take the mock test. They were also asked to immediately report any problems that they noticed, such as disconnection from the network, page loading taking too long, or pages not loading at all, etc.
We ran one test with default radio configurations, a second test with high-density settings, and a third test with high-density settings and DAS (Dynamic Airtime Scheduling) enabled.
We repeated the second test in the gym but this time with DAS enabled.
Finally, we tested video performance in the same high-density environments by gradually increasing the number of devices that were streaming videos.
The rules: Just like real life
Just as in a real-life setting, we configured static channels to make sure that conditions stayed the same for all of the tests.
At a random time, and without any announcement, we unplugged one access point to observe the impact of the possible AP failure during the critical examination period.
RF Environment: We are not alone
Healdsburg Junior High School is located in a residential area and we discovered an additional 60+ wireless networks (interestingly, almost all of them XFinity) that showed up in the cafeteria and gymnasium.
Test One was conducted using default Aerohive radio configurations. Students did not report any problems. Most of the iPads were connected to 5GHz and were not evenly distributed between the access points. Average client throughput was 3.89 Mbps.
Test two was conducted using the Aerohive high-density radio configuration, as recommended in a configuration guide. Students did not report any problems and client distribution between access points was dramatically improved. Average client throughput was 9.08 Mbps, which is more than 300% increase.
Test Three was conducted with high-density settings and DAS enabled. Throughput again increased dramatically, while the client distribution between access points remained extremely balanced, and students did not report any problems.
Failures for Tests One, Two, and Three
Two client devices lost network connectivity. We had to manually reconnect the clients and restart the exam application on these two iPads.
High-density video test
DAS was enabled while performing a video test where we simultaneously connected 150 streaming devices before encountering any issues. We then increased the device count to 180, and only a few devices disconnected or needed to be reconnected. When the load was increased to around 210 client devices, multiple devices were disconnected, while others needed to be refreshed or reconnected.
Is your network properly designed and configured for high-density?
Reading the results of multiple stress tests cannot answer this question. There are many variables and environmental conditions that have a profound impact on how a wireless network performs. All of these variables need to be considered when designing your network.
Designing a network using advanced enterprise-class configuration and tuning features, monitoring and insights with actionable intelligence, paired with intelligent troubleshooting tools and operational excellence, is your best bet.
K-12 Education Deployment Guide, Parts 1-5
These are the first five parts in a series of guides put together by Aerohive about setting up devices in a typical K-12 school environment. Each part is a fairly short topic that develops an extended example running throughout the series. Along the way there are numerous tips, suggestions, and alternative approaches. You can begin with the first five parts, which are ready now:
Part 1: Planning: This guide covers deployment planning, from setting coverage and capacity goals to conducting predictive and manual site surveys. It concludes by introducing the plans for an example high school deployment using AP230, AP130, and SR2148P devices.
Part 2: Wireless Access: In this guide, a network policy with three SSIDs is described. The guide shows how the SSIDs help satisfy the coverage and capacity goals introduced in Part 1.
Part 3: Ethernet Switching: The configuration of the network policy started in part 2 continues here. It explains how to configure an SR2148P device template and multiple port types to support various types of Ethernet traffic.
Part 4: High-density Radio Profiles: The high-density radio profiles to support administration of large standardized online exams in the school cafeteria and gym are described in detail. These settings were validated through testing at the Healdsburg Unified District Junior High School. A summary of the test results appears at the end of the guide.
Part 5: Bonjour Gateway: Bonjour Gateway enables clients to discover and access services on devices in different subnets/VLANs. This guide explains how to set up Bonjour Gateway so that teachers, staff, and students can access Apple TVs and printers in a different subnet and VLAN from their own.
All Posts in this Series