I should add that we did use lab simulation tools to design the network’s AP placement, using Ekahau’s WiFi scanning and design platform, which costs roughly $5k/year to license.
This approach is pretty standard practice in WiFi design: you upload building drawings into the design tool, identify major signal impediments such as walls, low ceilings, aluminized windows, and large metallic RF-reflectors such as baggage carousels. You then use the scanning component to walk the space and pick up and locate all potential interfering legacy APs. The tool lets you specify all construction materials with known attenuation values. Then you start placing APs on floor plans (these tools have built-in encyclopedias of all major APs and antennas, and knows all their salient characteristics), and generate simulated heat maps, making adjustments after each placement round. These tools also will predict “goodput” and user capacity, and a slew of other metrics. In our case, we discovered that placing many APs close together and at ground level (inside millwork and kiosks) let us exploit “people attenuation” to limit cross-channel interference with high-density crowds. And yes, Ekahau can simulate people as an environmental interferer. This approach gets you pretty close to the predicted network performance, but when you’re done deploying hardware, the customer typically requires demonstrating the system can reach the designed capacity. This is where Ixia came in. Before choosing Ixia we looked at several other solutions, including bread racks full of Raspberry Pi’s and an amalgam of lower-end test gear. Ixia ended up being the best bang for the buck. -mel > On Aug 26, 2021, at 7:19 AM, Mel Beckman <m...@beckman.org> wrote: > > I’ve used Ixia on high-confidence projects where we had to prove capacity of > an as-built network. Such testing isn’t cheap, but it’s sometimes the only > way to get the job done. > > Although you can buy Ixia gear and use it in a lab environment, that kind of > testing has limited application, because you often can’t fully replicate real > world external circumstances. > > For a Bay-area airport deployment of several hundred access points, I > specified and used Ixia’s Wireless testing platform to verify that we could > stream Netflix to several thousand mobile devices simultaneously. The Ixia > rigs were made mobile on carts, and each could simulate several hundred > simultaneous users. We hired this entire job using Ixia’s own engineers, and > it cost about $10K per day for their engineering labor and renting equipment, > over several days. > > We had to do this simulation in the as-built network ahead of the > Thanksgiving holiday, the busiest traffic day at any airport, as part of the > proof-of-capacity deliverable. There was no way to prove capacity in a lab > environment, due to the many unpredictable variables, such as access point > placement, Wi-Fi interferers, and back bone congestion. Another variable was > the Internet uplink, which consisted of two 5 Gbps BGP links to two different > providers. Even with this equipment, it was impossible to test the entire > airport terminal in one go. We did it separate test for the gate areas, and > baggage claim, the highest measured device demand locations based on the > previous Wi-Fi deployment. > > This was the only way to prove the network would not fail in the heat of > battle. I have to hand it to Ixia: their people were efficient and > professional, and knew what they were doing. There was no way that we as > network integrators would gain enough expertise to do this testing in such > short order. > > -mel via cell > >> On Aug 26, 2021, at 6:08 AM, Joe Yabuki <joeyabuki...@gmail.com> wrote: >> >> >> Hi all, >> >> I just wanted to know how you do your network testing when validating a new >> design/technology in your Network, especially to ensure that it will meet >> your SLA requirements for example that a voice call will not be dropped in >> case of a network element failure ? >> >> Do you test with IXIA, multiping, launch somes VM using ping with -i option, >> Windows ping by setting the timeout interval, or may be directly from the >> Network device (routers...), >> >> Many thanks, >> Joe
