Ryan is a student at Drexel University studying Mechanical Engineering with concentrations in Aerospace and Energy, and Systems Engineering.
Why is Computer Speed So important?
The current COVID-19 pandemic has forced businesses, universities, individuals, and pretty much everyone in between to take a more digital approach to their daily tasks and operations. Due to this new adjustment to lifestyles, a new emphasis has been put on the performance of technology, and on computers in particular. For students, this means more online classes. For businesses, this means more digital meetings. For engineers, there is more simulating and modelling. However, one thing that all of these fields have in common is that they are now more prone to connection issues, and face larger consequences when they happen. The objective of this experiment is to test what factors impact internet speed on a Dell G5 laptop.
The experimental methodology followed a straightforward, yet specific path. First, a browser was selected for the test, and opened. The only tabs that were opened in the chosen browser were a data collection sheet, Hulu’s website, and a tab that would use Google’s “Speed Test”. Next, the fan was manually turned to its lowest setting, before selecting a show. Once a show started streaming, in this case “Rick and Morty”, the download speed was recorded. After the initial test, the fan was switched on to its highest setting, and the internet speed would be tested again. Next, the initial browser was closed, and a second one was opened. The same tabs were opened, and the procedure was repeated, and the data recorded. Once all browsers had been tested, the ethernet cable was removed, and the computer relied on WiFi signal for internet connection. The test was then repeated, and the data recorded.
This test was repeated five times, and spread out over a week so as to avoid over-using the computer. Additionally, the test was performed early in the morning, before anyone else would be using the WiFi at the same time (namely, roommates). This was to prevent any inconsistencies in how the connection was being used at that point in time.
Given the details about this project, certain assumptions could be made while testing. This experiment would run a total of 120 tests, allowing for normality to be assumed due to sample size. Additionally, the factors of the project are all independent and do not affect each other. Each one is normally controlled automatically by the computer, but is able to be overridden manually by the user. Therefore, each factor is independent of the others.
In this experiment, it was hypothesized that the download speed is independent of the chosen factors - in other words, no matter what conditions the computer is under, the download speed will not fundamentally change. The alternative hypothesis was that at least one factor affected internet speed.
Upon completion of the experiment, data was recorded in SPSS, and a multi-factor ANOVA was performed. Using a confidence value of 0.05 (5%), the significance of each factor and their combinations was calculated. Connection had a significance value of 0.000, Streaming had a significance value of 0.025, and Fan had a significance value of 0.028. These all fall below the 0.05 threshold, and the hypothesis can therefore be rejected. Browser had a significance value of 0.424, and is the only individual factor to be over the 0.05 threshold.
In combined factors, “Connection*Streaming” received a significance value of 0.039, while “Connection*Browser” and “Connection*Fan” had values of 0.052 and 0.162, respectively. “Browser*Fan” received a value of 0.020, and “Browser*Streaming” was given a 0.092 value. “Streaming*Fan” received a significance value of 0.556.
SPSS calculated values of 0.288 and 0.049 for “Connection*Streaming*Fan” and “Browser*Streaming*Fan”, respectively. “Connection*Browser*Streaming” received a significance value of 0.000, “Connection*Broswer*Fan” received a significance value of 0.004, and, finally, “Connection*Browser*Streaming*Fan” had a significance value of 0.001.
A Post-Hoc Tukey’s Test was also performed on the Browser factor, and analyzed how much difference there was between each level. All browsers were grouped into the same subset, indicating similarity in their impact (or lack thereof) of the internet speed tested.
Analyzing the data showed that the hypothesis, that download speed is independent of the listed factors, can be rejected. It was found that the connection (Ethernet vs WiFi), whether or not the fan was on low or high, and whether or not the computer was streaming all impacted the computer’s internet speed. Of the factors listed, connection had the highest influence on the computer’s performance. On contrast, the sample provided suggests that the chosen browser (Edge, Chrome, and Firefox) has no impact. The Post-Hoc Tukey’s Test determined that each level is similar enough to the others that they can all be grouped in the same subset.
The largest issue faced was the fluctuation of internet speed when being tested. At any given moment, a computer can oscillate between ranges of over 100 Mbps in seconds. To combat this, multiple replications of each test were performed. Acquiring more data made it easier to help identify outliers in the results, and allowed for an average response to be calculated.
Scalability can be applied for this project, as it tests the technology in the computer. With a larger sample size, it is expected to keep the same trend. The actions required for the test also do not test the computer enough to wear down any of the computer's technical abilities, and can be repeated a large number of times without failure (assuming no external factors). However, that is limited to the Dell G5 laptop. When expanded to other pieces of hardware, new tests must be performed to ensure accurate results.
There is a quickly rising importance placed upon computer performance in the age of COVID-19. This experiment was designed to determine what factors impact internet speed. When certain levels of each factor are combined and the internet speed tested, a data sample is created that gives an idea of how the computer is affected under certain conditions. In this experiment, a Dell G5 laptop was used as the medium for the tests. It was determined that connection, the fan speed, and whether or not the laptop was streaming altered the internet speed of the computer. The chosen browser, on the other hand, does not affect the computer performance. Similarly, the different browsers provide minimal difference between each other. Therefore, the hypothesis that none of the factors affect computer performance can be rejected.
© 2021 Ryan Clancy