With the COVID-19 pandemic, we are facing an unprecedented situation on a global level. Entire countries are in lockdown, and people everywhere are being asked to stop or minimize their physical (social) interactions. Our broadband connections are becoming our lifelines – figuratively and literally: we are using them to get news, connect to our work environments (now all virtual), and for entertainment too.
The wave of special emergency measures continues across countries and continents. Let’s take a look at the first effects of this global pandemic on network traffic. In a way, this may be the “shape of things to come” for all other networks.
We analyzed data from several networks in Western Europe from the week of March 9, 2020. We looked at the traffic volumes and patterns. While we were still in the early days of the new reality introduced with special measures, we were able to spot some network anomalies and trends.
Some countries in Western Europe have declared lockdown measures – Day 1 in our analysis is the day the lockdown was enforced. We also continued looking at the situation over the following weekend after the lockdown was declared – which we refer to as Weekend 1 (or Sunday 1, the Sunday in Weekend 1). To provide a better comparison, we looked at day-to-day (workday and weekend day) comparisons against the previous week.
Social media and messaging
On Day 1 and Sunday 1, we saw a significant increase in popular messaging and social media applications. One of the most popular, WhatsApp, exhibited an increase of 117–217% during Day 1, with apparent spikes in the morning hours on the first day of the school lockdown. We also saw a sixfold (or 500%) traffic increase on Sunday 1. Even with these spikes, the overall WhatsApp traffic (in Gbps) remained manageable in both cases – from the perspective of the total traffic volume in the network.
Netflix
On the online entertainment front, the increase in Netflix traffic showed that people started streaming earlier in the day (morning hours and early afternoon hours). While the evening traffic volumes remained within a +20% range compared to typical values, the overall increase of traffic during the day ranged from 97% (morning) to 27%–42% (early afternoon), significantly contributing to the total network traffic.
A look at the Netflix traffic on the weekend brought much more concerning stats. Volumes rose between 54% and 75% (in peak viewing hours) compared to the previous weekend, with substantial contributions to the overall traffic. Knowing that weekend evenings are the peak traffic hours for most networks, Weekend 1 brought more stress to the network – in relative and absolute terms (terabit range).
Total network traffic
Day 1 brought an increase in the overall traffic of anywhere between 30% (5 p.m.) and 80% (9 a.m.), with boosts between 50% and 70% for most of the day.
The Weekend 1 Sunday traffic rose between 34% and 97% during the day, bringing additional terabit-levels of traffic volumes to the network at times. While “normal” weekends (because of a lot of online streaming and gaming) can be used as a solid baseline and benchmark to check whether there is enough capacity to address these peaks, it seems that Weekend 1 additionally stress-tested the network and in a significant way. The good news is that networks were able to absorb this peak without degrading the quality of service. The bad news is that this may be a new trend. The future will tell.
Final thoughts
What conclusions can we draw from Day 1 and Weekend 1 snapshots? What are the new broadband realities that we may be seeing, locally and globally, in the coming weeks or months?
Here are a few thoughts.
We are witnessing an increase in traffic in both wireless and wireline networks. However, with prolonged times spent at our homes in the lockdown periods, wireline networks will play an increasingly important role. Service providers need to ensure that their internet infrastructures are up to these new tasks – with enough capacity and ability to deliver all services with high performance under the increased traffic demand.
Service providers need access to real-time and granular information about their networks. They need tools that can help them gain deep insights and correlate and visualize data at speeds and in ways that human brains cannot. A new generation of big data-driven network intelligence and analytics technology such as Nokia Deepfield can help obtain real-time, very granular (down to application level), accurate, multidimensional details about the network, services, network flows, and usage patterns. (By the way, this is the technology used to analyze all data points used in this article.)
While we will continue to monitor the global and local effects of the COVID-19 pandemic on networks worldwide, it is crucial that we “keep our cool.” The lessons we are learning and the challenges we are facing will forever change the way we work, communicate, and entertain. Telemedicine, for example, will be one of the areas where these lessons may bring the most profound changes to how health services are accessed and delivered.
A comforting thought is that we are at a safe place, technologically speaking. With the latest technology advances and standards, we have ubiquitous, sophisticated connectivity – from our home networks, across service providers’ and hyperscale realms. The newest networking technologies are ready to rise to the challenges brought in front of them. However, in addition to continuing 5G and fiber access infrastructure rollouts, we may need even more investment in the networks (wireline, wireless, and converged) to make them better and readier for the future.
That will hopefully be a more normal-looking future, which everyone wants to see again.
Share your thoughts on this topic by joining the Twitter discussion with @nokianetworks or @nokia using #bigdata