In Part 1 of our Buffalo Connect analysis, we touched upon some of the basics of the network. One of our central topics was covering the speed of the network and how it could be used successfully (or not). Today, we're going to be going a little more in depth into the speed, the network's signal strength, and routers.
Let's get started...
WiFi technology has been an evolving standard. It's important to understand a little about how to quantify the speed of a WiFi network before actually going into it.
The four main WiFi standards we'll make note of are 802.11b, g, n, and ac. Out of these four, b came out first in 1999, g came out in 2003, n in 2009, and ac in 2014. If you've ever bought a wireless router for your home and searched to see if it was a/b/g/n/ac (ordered by increasing speed), then you have the general concept. Below are a couple charts that detail their respective distances and speed potentials:
Broadcast distance of wireless signals has continually increased through 2009. But even though the actual distance of signals did not increase with the release of the 802.11ac standard in 2014, the actual speeds transmitted over those distances did. A router that has been upgraded from n to ac generally sees about double and sometimes triple the speed when the wireless device is the same distance away. This is important to note when we go into further detail about the routers used throughout Buffalo Connect further in this post.
In short you can get a good idea about how much data transfer rates and signal distances have increased in just the past decade or two. While these speeds and distances are the physical maximums--and real world situations can cause these numbers to vary--the evolution of wireless technology has continually improved.
In order to visualize the corridor, we mapped out the download speeds at various points across the corridor using a nifty app called NetSpot (NetSpot Wikipedia Page). NetSpot is great for collecting data about wireless networks, and it allowed us to generate some very clear results:
You can see above that we recorded download speeds anywhere from 0Mbps up to 4.1Mbps.
But wait a second--didn't we say that the highest speeds we recorded on the network didn't exceed 1Mbps? Then where did 4.1Mbps come from? Although we have reason to believe that the cap has now been increased from 1Mbps to 2Mbps, that still doesn't explain the 4.1Mbps we're seeing.
While doing our testing with NetSpot, we also ran some Ookla speed tests, which showed initial bursts for the download speed of anywhere between 10-20Mbps. The bursts then quickly, but not immediately, dropped down to the speeds that we were seeing when we initially ran tests back in August and that were in line with general use.
(In our next post in this series we'll analyze the burst by exploring more of UB's infrastructure by comparing their network for students, faculty, and staff versus those on campus who are just provided with guest access.)
This led us to think that from that initial burst we're getting a glimpse of the true power of the network, and then it is subsequently capped down to 2Mbps for both upload and download. Although there is typically an initial burst in speed when transferring something, the extensive burst we experienced is indicative of something a little more.
Getting back to the point, Netspot's tests would stop before coming down from that initial burst to accurately reflect the new 2Mpbs cap.
Here are several of the Ookla speed tests that led us to this conclusion:
We can infer from both the graphs up top and from these speed tests that the current speeds offered via Buffalo Connect were in line with technology available nearly two decades ago.
Is the city actually using tech from twenty years ago? Or was the roll out of Buffalo Connect done with cost saving measures, considering there doesn't seem to be much practical use for the network?
Signal Strength/Signal-to-Noise Ratio
Now to analyze the Signal-to-Noise ratio. This is a measure that compares the level of a specific signal to the level of background noise, e.g. microwaves, other WiFi networks, and Bluetooth devices. Basically it's the reliability and consistency of a network. The Signal-to-Noise ratio is defined as the ratio of signal power to the noise power, often expressed in decibels. This value is not expressed as a true ratio, but rather as the difference between the level of noise and the signal floor.
The Signal-to-Noise ratio is a very important measurement because without reliability or consistent throughput a network doesn't exist. What good is a network at higher speeds if the connection continues to cut out? Or in Buffalo Connect's case, what is the point if you have a reliable network but traffic is at a snail's pace, like cars in a traffic jam when the speed limit is 65?
Take a look at what we found:
The map here might look a little messy, but it tells us the reliability and strength of the signal from Buffalo Connect drops off quickly once you get off of Main Street.
Drawing from this key, we can see there's a very fine line of green teal down Main Street, and pretty much everything else is blue. The areas that aren't on Main Street are generally blocked by large buildings.
While on Main Street, there is a lot of background noise. Yet with all the background noise, it still has a strong signal. If you have potential for decent download/upload, but a lot of noise and/or interference, then you will likely experience slow or "unstable" connectivity that appears to drop. While on the Buffalo Connect network, while going up and down Main Street (the green areas on the map), and throughout Canalside, the connection consistently provided 1Mpbs and there was no drop in coverage.
When today's technology can broadcast signals at least 800 feet (probably more, considering this number actually represent the n standard, as the numbers for ac aren't reliably established) around each router, the network should not be limited to just specifically Main Street and Canalside. With that reliability, they should broadcast much faster speeds at greater distances than we're seeing.
This may seem like a contradiction to what we were saying about the network using technology that's over two decades old, but it's not. While the speeds and distance are on par with standards established twenty years ago, the Signal-to-Noise ratio indicates that this tech is capable of much more modern speeds.
To test this theory we further investigated the network's hardware.
We took some pictures of the routers to see if we could determine what M&T/the City decided to purchase. We weren't certain we'd be able to pinpoint exactly what we were looking at, but thanks to some good luck and good sleuthing, we've determined Buffalo Connect utilizes very modern, and expensive, routers:
These are Cisco Meraki MR72 802.11ac Outdoor AP routers. The base model retails for around $1,200 per router. We're hoping that since dozens of dozens of these were purchased, they received a bulk discount. Regardless, a pretty penny was spent on these brand new routers, all to provide residents, businesses, and visitors to the area with a wireless network that we have trouble finding any practical value for.
On top of routers, there are other aspects to the network that would've cost the provider more money, such as the black boxes pictured behind the router, the lines to the router/black boxes also pictured, the lines from UB's fiber, labor, weather-proofing, setting up firewalls, and basic network administration. While we'd have trouble estimating the total costs of all this, we can safely assume it wasn't cheap.
To reiterate our thoughts from Part 1, Buffalo Connect is a lackluster network. As one can gauge from our further analysis, it has a lot of potential to be greater than it is. In Part 3, we're going to investigate UB's network to see how it influences Buffalo Connect's performance.
authors: Sanjay, Quinn, Cola'