Welcome to the fourth installment of Preseem’s Know Your WISP Network blog series! Over the past few weeks, this series of blogs has applied the insights contained within Preseem’s Fixed Wireless Network Report to demonstrate the real-world experience of fixed wireless subscribers, networks, and equipment. In our last entry, we explored how WISP access point (AP) throughput rates compare across the AP market and different AP models. In this blog, we’ll again explore metrics related to APs, with a particular focus on latency.
As a result of our analysis, the WISP AP latency insights that we derive will answer the following questions:
- What amount of latency is incurred by APs across the market?
- How does latency compare between AP models?
Continue reading to get the answer to these questions. When you’re done, click here for a free download of the entire 26-page Fixed Wireless Network Report!
AP Market Latency
We’ll begin our exploration of latency with a broad analysis, looking at access points across the market. To accomplish this, we compiled data from the billions of metrics that Preseem compiles every day from WISPs across the globe. After examining Preseem’s data pool, we derived the results presented in the figure below.
Interestingly, the distribution of latency across all AP:s presents a peculiar pattern. Most access points deliver service with less than 100 ms of latency during peak times, yet a significant number in our study are over that benchmark. But what is a “good” latency value anyway?
Well, it’s common knowledge that latency requirements differ a lot from application to application. For example, high latency has very little effect on Netflix but has a large impact on gaming. However, a simple point of comparison is Voice over Internet Protocol (VoIP). Typically, the end-to-end latency for a VoIP call must be less than 150 ms for the user to have a decent experience. Looking at the data in the above figure, we can see that the vast majority of APs are under this threshold.
Note that the latency values presented in this analysis reflect the latency experienced from Preseem to the subscriber and back. Therefore, these latency values do not include the rest of the network path. This explains why the latency values must be lower than 150 ms to achieve a good VoIP experience.
AP Model Latency
Let’s now narrow our analysis by looking at how latency compares across different AP models. The figure below maps the latency of various AP models within the 80th percentile.
As you can see, there’s significant variation in latency across AP models. Also, the results indicate that there’s a general trend in which newer AP models have better latency characteristics compared to their older counterparts.