Network Site Surveys
By: Abhishek Shah, Alex Zhou, Kexuan Liu, Marcus Amal, Ivy Wang
Our project
In today’s technology driven world, connection to the internet becomes increasingly important, to the level where most people can’t imagine a life without access to the internet. Yet many underprivileged people live that reality. Affordable internet access should be given to as many people as possible, and our team hopes to enable that.
Background
Our team was formed from a team of students at UW as a part of the Community Networks capstone class. The class is directly connected with SCN, Seattle Community Cellular Networks, which strives to provide underserved communities with affordable internet access. Instead of paying and using the internet from standard ISPs, SCN wants community members to participate in a mesh network, where the network traffic can be routed in a variety of ways. SCCN is open to anyone who is interested in the project, and does not require any technical knowledge.
Thanks to community outreach, SCN puts cellular network cores at various locations that form the basis of mesh networks. Two new sites are Garfield High School and Franklin High School. Network strength is highly variable because it can be affected by a wide variety of factors, like buildings being in the way or the angle of the receive being off slightly, so it is important to test the network’s feasibility by actually going around and measuring network speeds manually. Our team was on the task of doing these site surveys: collecting data, organizing it, and presenting it in clear fashion.
Design
To test the quality of the network in Garfield High School and Franklin High School communities, we actually went to those two places and did the two site surveys. For each of those two sites, we received a list of different places in that community and we tested the signal strength and speed of the network at each of those places. After getting the network quality data at each place, we organized them together in a file and visualized them to show a more clear overview of the network quality across the whole community and help stakeholders better understand the trend of the network quality (around which place the network connection is bad or good).
Implementation
We collected the data of the network quality of both a CPE and a cell phone. We collected the signal strength and speed test results (ping, download and upload speed) three times using a laptop connected to CPE and a cell phone connected only to the network. Specifically, we used Bai Cells CPE for both site surveys. However, for Garfield High School site survey, we used the CPE with model EG7035L-M11 and for Franklin High School site survey, we used the CPE with model EG7035E-M11, which has a higher gain and better performance. We used a very long bamboo pole for Garfield High School site survey and a relatively shorter foldable metal pole for Franklin High School site survey. We also used a battery bank to power the CPE as we moved from one place to another.
On the computer, we used the information shown on the overview page to collect the signal strength of the CPE and used the online Google speed test to collect the ping, download and upload speeds of the CPE. On the cell phone, we used the Network Cell Info Lite app to collect the signal strength and the SpeedTest app to collect the ping, download and upload speeds. We did all of those tests three times and put our data in a Google Sheet. Then, we used Google Map to visualize those data points.
For each of the two site surveys, after creating the map, we shared the editable version of it with SCN and also printed a preview of it which is included in the “Reflection” section below.
Challenges we faced and how we overcame them
● After Garfield High School site survey, we found that the CPE used in that survey might not have enough gain to get a good signal. We overcame this by using a higher gain CPE in the next Franklin High School site survey.
● During our fieldwork, we found that the signal strength varied a lot, even of two locations close to each other. We overcame this by paying extra attention to the direction of the CPE and testing more places in the next site survey.
● Transporting equipment was tricky. Specifically, the poles that we attached the CPEs on were very long and heavy. For the unfoldable bamboo pole, it could only be transported via truck, so there was additional planning of figuring out how to get the poles around. We overcame this by using a foldable pole in the next site survey which was lighter and much easier to transport.
Reflection
Result:
The screenshots above are basically what it looks like. Each star icon is the place measured and the yellow home icon is the signal source. Color is based on the signal strength. Click on each icon then it will show the name of the place and the signal strength there.
(green) signal strength -90 ~-80 dBm
(yellow) signal strength -100 ~-90 dBm
(orange) signal strength -110 ~-100 dBm
(red) signal strength ~-110 dBm
We collected data, organized it, and presented it in the above visualization. We think this visualization can provide information for the network’s feasibility around Franklin and Garfield sites.
Analysis:
As we can see from the above visualization, signal strength is highly correlated with the surrounding buildings between the signal source and the locations. Places where we can see the signal source clearly have a much better signal than places where we can’t see the signal source. This means that when there is no building blockage between the test location and signal source, we will have a good signal. We got a great signal in Lake People park because this park is a park on a small hill and we can see the roof of the Franklin high school clearly in this park. The influence of the building blockage was huge and surprised us. As we can see from the Franklin visualization, College Possible Washington and Baked from the Hart are just two buildings facing each other on a road, but College Possible Washington got no signal. This makes the site survey important because the signal strength varies a lot and we need to actually go around and do the measurements.
Our goal is to help the Seattle community network to meet the network needs of the community by providing an affordable network to the community. This user site survey can help us have a good understanding of the current network strength around a certain area and the general topography and geography of the area where we install our networking equipment. The location of user sites depends on who we want to focus on providing the network to, which tends to be lower income members of the community like Nickelsville. After getting data from the user survey, we can know how well our current network works and decide how to improve it based on the current signal strength and geographic information.
Next step:
In the user site survey, we collected data and did the visualization manually. We repeated the process in every location and site. Then a visualization tool that can take in data from excel and make the visualization would be very helpful since in order to know the network strength of different sites we need to do the site survey on multiple sites.
Also, we noticed that the signal strength is greatly affected by the surrounding buildings, so it is good to not only do a 2D map visualization, but also a 3D visualization to see buildings or hill blockage. This will be helpful to determine why we have a signal strength distribution like this and decide how to improve the signal strength for a certain place because we can figure out the reason why one place has a bad signal. In our site survey we find out that one location has bad coverage because it is actually behind a building, while another location down the block has a strong signal because it has line of sight. Thus, a 3D visualization can provide more information for the user and help the community network improve.
The next step for our project might be to make a visualization tool that can handle the data efficiently and support the 3D visualization. Then every team who needs to do the site survey can use this tool instead of doing the visualization manually and slowly and they will also understand the data better with the 3D visualization.
Advice:
The User Site Survey that takes the actual measurements was a pretty clunky task and needed to be conducted by a team. It is good to have at least 3 people in the team: one person needs to carry the pole, one person needs to hold the computer and the last person is responsible to do the cell phone signal test and record the data. Also, transporting equipment has also been tricky. Specifically, the poles that the team needs to attach the CPEs on might be difficult to transport. If the pole is very long and unretractable and can only be transported via truck, it is good to have an additional plan to figure out how to get the poles around.
Ideally, to make user site testing more accessible, the team can find a collapsible flag pole.
Since the antenna is directional, the team needs to be careful when they try to test the CPE signal strength. Direction really matters. A slight rotation of CPE can cause a big difference in signal strength. Ideally, the team needs to let the CPE point to the source antenna. It is good to try multiple directions when the team can’t see the signal source. In one location test, the team needs to keep track of the CPE signal strength, CPE speed test, Cell phone Signal Strength, and the Cell phone speed test. All of these measurements have to be repeated multiple times, and organizing this data can be difficult if you don’t have a plan beforehand.
credits + contacts+ references
Interests:
Abhishek: My interests outside of this project are security and systems level protocols for optimization. I am currently working on a research paper to create a new OTP protocol that adds a layer of encryption to the entire process and hides the plain text codes from telecoms and any other malicious adversaries.
Kexuan: My interests outside of this project are software development and database management. I have done many interesting projects in my classes related to those two fields.
Ivy: My interests outside of this project are software development and data visualization. I think data visualization can help people understand the data better and also be interesting.
Marcus: My interests outside of this project are spending time outdoors, eating food, and hanging out with my dog!
Alex: I’m a big fan of longboarding and skateboarding. In my spare time I also play video games and occasionally go climbing at the IMA. One ongoing project I have is a novel I’m working on that is a sci fi book about identity.
Reference:
Many thanks to Esther Jang and the ICTD lab.