How ISPs and operators ensure network quality in a pandemic
COVID-19 has caused a spike in remote workers, increased internet usage and traffic surges. Here's how ISPs and telecom operators plan for continuity and quality during the pandemic.
The internet is fine during the COVID-19 pandemic, or maybe it isn't. Work from home is working, or maybe remote workers can't even call in for a conference. How is the internet really doing, and what are suppliers doing to ensure network quality and support traffic in these exceptional times?
In this article, we discuss short-term tactical responses, middle-term measured responses and long-term strategic responses internet service providers (ISPs) and telecom operators take to maintain continuity and network quality in a pandemic.
Internet core strategies
Internet usage has always been bursty, and suppliers responsible for the various pieces that make up the internet don't want to be the weak link. The reasons are simple: Customers don't know where an internet problem comes from, only that they're having one; access operators are quick to end relationships with weak partners; and users are even quicker to change service providers or access ISPs.
Tactical response: Use residual capacity. The first measure providers take to ensure network continuity is to build out capacity in the internet core, content delivery networks (CDNs) and search engines. It's cheaper to provide capacity than to lose customers, and the tactical response to the pandemic is simply to use this residual capacity. This isn't a response as much as an ongoing strategy.
Tactical response: Encourage streaming providers to reduce bit rates. One tactical response operators use to respond to increased traffic is to encourage streaming providers to reduce their bit rates, either by better compression or by delivering a lower-resolution video.
Operators that have prepared for capacity problems can also downsample video -- dynamically reduce resolution -- especially for mobile delivery, to reduce bandwidth requirements. Some operators, particularly mobile operators, may throttle data and download speeds if regulators and contracts permit. This hasn't happened much in the current pandemic crisis; in fact, most major providers are temporarily lifting their throttling caps.
Measured response: Improve peering. Even with a lot of capacity, the internet is vulnerable to traffic, so the Peering portion of the internet is designed to maintain a series of paths from each user to each possible destination. These paths are monitored constantly, and traffic is routed to an alternate path if congestion or failure occurs. This traffic rebalancing, combined with the substantial capacity built into the internet's core, has enabled the internet to ride out the pandemic traffic boom and deliver good service. This improved peering, including peering with CDNs, is a measured strategy providers take.
The internet is only as good as access, and access comes through wireline ISPs or mobile networks.
Strategic response: Use high-speed optical transport. ISPs say they're beefing up their capacity by accelerating their expansion plans for their core networks -- the federation of operators and networks we call the internet. Even if ISPs didn't expand their core networks, technical indicators say the internet is handling things fine. Providers' strategic response to internet assurance is to use higher-speed optical transport, up to 800 Gbps per fiber.
However, many people don't believe the internet is responding well to pandemic-related traffic surges because the service quality potential of the internet core isn't the only issue. The internet is only as good as access, and access comes through wireline ISPs or mobile networks. Both these mechanisms expose a problem with the economic digital divide.
Access network strategies
The majority of network cost isn't the core; it's the access network. Wireline network providers must justify the last-mile costs for each customer's revenue. Even with paths that combine fiber, CATV or copper to create a higher-level trunk that heads back toward the internet, fewer customers share the cost of that trunk than those that share internet core costs.
This means little to no excess capacity is built into many access networks. ISPs' ability to overbuild for capacity depends on customer density and the value of services they purchase. As a result, rural and poor areas will likely have less capacity to absorb new peaks of traffic. They'll also have fewer facilities that provide redundancy for routing around failures or congestion.
The problem is particularly acute in areas where the only wire in wireline is the twisted-pair copper loop. Digital subscriber line (DSL) technology can deliver broadband across the last-mile loop, but speeds will typically be 25 Mbps or less, which is marginal at best even for normal streaming video. Congestion is likely to happen anywhere in a DSL-based access network, so DSL customers are more likely to experience problems during the pandemic than those on fiber or CATV cable.
Strategic response: 5G and fiber to the node (FTTN). ISPs have few good near-term responses to the wireline access problem. The big hope here is that 5G technology -- including a combination of millimeter wave 5G and FTTN -- will reduce the cost of delivering wireline broadband, where fiber to the home or CATV cable isn't economical. This is the strategic response to wireline access limitations on network service quality.
Tactical and measured response: Upgrade VoIP servers. Mobile and VoIP services have their own access problems. Voice calls rely on a set of servers -- Session Initiation Protocol (SIP) servers, for example -- that translate between phone numbers and the IP addresses needed for internet routing. These servers can overload even with plenty of available internet and access network capacity, and multiple SIP servers are in the path of calls, often within companies themselves.
Many remote workers have experienced a "no circuits available" message when attempting to make a voice call during the pandemic. This problem comes from their own company's VoIP servers and call handling. VoIP servers are relatively inexpensive, so providers and companies are already upgrading these servers to handle the load. These upgrades are both a tactical and measured response to load issues. However, the majority of voice call problems reported during events that drive companies to implement work-from-home polices are caused by the companies' own VoIP gateways.
The specialized mobile network devices, the backhaul fiber and the cells have capacity limits that the pandemic can exceed.
Mobile network strategies
The mobile network also contains specialized devices for mobility management and phone registration to track users as they turn on their phones or move between cells. The cells themselves must connect back to the mobile network -- cellular backhaul -- usually through fiber optic trunks. The specialized mobile network devices, the backhaul fiber and the cells have capacity limits that the pandemic can exceed. These limits make mobile networks especially vulnerable, particularly if mobile users aren't at home or don't use or have Wi-Fi calling.
Mobile operators have been moving away from special mobile devices for registration and mobility management, such as devices for IP Multimedia Subsystem and Evolved Packet Core, respectively. Instead, they are hosting software on servers for improved scalability and resilience under extreme loads. Operators make this transition when subscriber numbers and subscriber revenue potential are highest. But, because mobile is more competitive than wireline access, mobile operators are accelerating their plans to use hosted mobile components to improve service.
This shift to scalable mobile network technology involves all three response strategies:
Tactical response. Operators can more quickly transition to hosted mobile components, where mobile elements are already software-based.
Measured response. Operators can plan additional hosting capacity for these mobile elements.
Strategic response. Operators see hosted mobile features as a better option for the future.
The longer-term response to the problem is 5G. While most of the 5G talk focuses on the potential for higher speeds to the user, the real benefit of 5G is greater total capacity per cell site. This means a 5G network could support more high-traffic users simultaneously. 5G's mobility management and registration -- the 5G Core -- is also based on scalable, hosted components rather than fixed devices, so a full transition would create a mobile network almost as elastic and efficient as the internet itself.
Lessons from the COVID-19 pandemic
The real lessons of the COVID-19 pandemic are simple. Telecom operators, cable companies and ISPs are spending on service quality to the extent that revenues and competitive pressures justify. Technology shifts -- including improvements in fiber optic transport and agile hosted network elements that replace conventional devices -- could enhance and ensure network quality in a pandemic but will be rolled out only as revenues and competitive pressures dictate.
So far, the COVID-19 pandemic has most greatly affected the areas where networks are most profitable and competitive. Those areas are likely to see network quality maintained through the course of this pandemic. If the virus spreads to more rural states and counties, quality constraints set by low user density and total revenue potential at the access level could reduce the quality of service, enough to widen the digital divide just when people need online access the most.
