ISPs rely on capacity for COVID-19 internet traffic spikes
As COVID-19 causes surges in internet traffic, internet and network providers have responded by increasing capacity and installing more fiber. So far, it's worked.
In the days of COVID-19, internet service providers, carriers and telecom operators are relying on available capacity to support the millions of people using the internet.
As nations and organizations across the world responded to the pandemic with stay-at-home orders and work-from-home policies, the internet experienced traffic surges and unusual usage patterns. The internet traffic spikes weren't anything new. But, before COVID-19, a traffic surge might have stemmed from thousands of users downloading the latest release of Fortnite, said Andrew Dugan, CTO of CenturyLink, during a USTelecom webinar on internet network performance amid the COVID-19 pandemic.
Internet providers have learned to respond to these internet traffic spikes and hotspots with traffic engineering -- a method in which they shift traffic to relieve congestion -- and capacity planning. They measure packet loss and latency to evaluate congestion and gauge the ability of links to handle traffic, Dugan said. This expertise has paid dividends in the current coronavirus situation, but the stakes are high.
"One of the biggest challenges in the pandemic is the sheer volume of traffic and the shifting usage patterns," he said. In the first couple weeks of the pandemic, various parts of CenturyLink's network infrastructure "ran hot," driving the provider to quickly increase capacity with fiber. Internet peak times strayed from the usual evening hours, and collaboration tools caused major jumps in traffic throughout the day.
"It created a slight amount of anxiety as traffic began to ramp up," Dugan said. "We saw a 35% increase in internet traffic, with some days bouncing around as certain events bumped or increased traffic."
One of the biggest challenges in the pandemic is the sheer volume of traffic and the shifting usage patterns.
Andrew DuganCTO, CenturyLink
Comcast Business saw similar traffic spikes, with upstream traffic surges ranging from 32% to 60% in large areas like Seattle and the Bay Area, said Noam Raffaelli, senior vice president of network and communications engineering at Comcast Business, during a company webinar. Video conferencing and Wi-Fi usage increased 228% and 24%, respectively.
Although these spikes were drastic in some places, Raffaelli said Comcast's networks are built to handle surges like this.
"For us, this is BAU [business as usual]," he said. "We are building and dimensioning our networks to handle these types of shifts in peaks." Still, his team has taken a short-term response to these shifts by augmenting their 100 Gbps links and installing additional fiber to absorb traffic increases, he added.
As more employees work from home during the COVID-19 pandemic, video conferencing and Wi-Fi usage has spiked.
Factors for successful internet network performance
Ten years ago, network infrastructure might not have been able to withstand the internet traffic spikes caused by the COVID-19 pandemic today. Over the years, however, providers built their networks and backbones using a similar formula with capacity as the common denominator.
Balanced peak internet traffic times. While more employees and students are using the internet from home, the usage occurs primarily during the day. But the daytime, traditionally, isn't the peak time for internet usage -- evening is, when people get home from work and start gaming or streaming, Dugan said. As a result, internet use has balanced out a bit more.
Raffaelli said Comcast Business has seen the peak patterns change as well. While traffic historically spiked around 9 p.m., the provider now sees downstream peaks around 7:30 p.m. and upstream -- or outgoing -- traffic peaks from 6 p.m. to 8 a.m.
Built to withstand failure. Both CenturyLink and Comcast Business emphasized how they build their networks with extra room for failure, considering capacity, resiliency and efficiency. By purposefully building leeway into their networks, providers can take a proactive approach to disaster.
"Fiber cuts and equipment failures happen more than anybody would like, so you have to build that into your architecture," Dugan said.
Planning rules and guidelines. In addition to planning for failure, internet providers also build redundancy into their networks because they don't want to come close to the edge of their capacity. "You tend to be months ahead because you don't want to run to full capacity," Dugan said.
Comcast Business, for example, works on a long-term forecast that assesses capacity needs 12 to 24 months out, Raffaelli said.
Traffic engineering. Network operators have policies and procedures in place to handle disasters, Dugan said. One such example is traffic engineering, which enables a provider to override congestion. "We've certainly taken advantage of this during COVID," Dugan said.
Challenges for internet connectivity in a pandemic
Although most providers can quickly control capacity in their private internet backbones, connectivity gets more complicated at the customer edge, which supports multiple connectivity types spread across millions of customer locations, Dugan said. Managing those connections grows more complicated in rural areas that rely primarily on DSL or cable.
"The physical medium as you go out to rural areas is typically DSL, but the physics don't allow you to run high speeds," Dugan said. "Where you have fiber optic systems, it's easier to manage at scale. You run into trouble when you have the slower, physical-based infrastructure."
Remote workers and users who depend on these copper-based connections will need to consider bandwidth constraints, especially if members of the household spend a lot of time gaming or streaming, he said.
"DSL and cable allocate the bandwidth to give more bandwidth into the home than going out," Dugan said. "Be aware that you may need to reduce usage, thinking about both upload and download directions."
