As mobile networks have become the dominant way for users to access the internet, traffic levels on mobile consumer services have escalated quickly. The Ericsson Mobility Report is a useful source of information to quantify this growth, and predicts user traffic on 4G networks continuing to grow strongly over the next few years. But with a 10-fold growth, it’s 5G networks that will likely carry the lion’s share of user traffic by 2028.
A key contributor to this growth will be fixed wireless access (FWA), where customer premises equipment (CPE) devices in homes and offices use mobile connectivity rather than wired. Up to 64 million (4G) and 236 million (5G) FWA connections are forecast globally by 2028, and FWA will be responsible for more than a third of global mobile traffic by then.
FWA has always been appealing to internet service providers (ISPs) looking to offer fixed broadband services in markets where there isn’t good last-mile copper or fiber available. What’s changed is that many ISPs are now also looking at 5G FWA to deliver high-speed broadband services in markets that have existing last-mile connectivity—especially where the existing infrastructure would require an expensive physical upgrade to deliver the speeds many consumers now expect.
FWA will bring the immense benefit of faster internet access to many communities, but from the mobile operator’s perspective, it comes with a lot of consumer IoT and other end-user infrastructure that has historically caused problems on wireline networks. This is a big change for mobile operators, especially in terms of managing threats.
Since COVID-19, we have seen mobile networks being used increasingly for gaming, with a corresponding rise in DDoS activity to, from, and between mobile users (this has been happening on wireline networks for years). And, as more consumer IoT is connected to mobile, there’s already an increase in compromised device populations (botnets) connected to mobile services. This drives additional growth in DDoS, scanning, and other unwanted activities.
So, what’s the issue? We’ve seen all of this on wireline networks for years. Well, mobile networks are different. There’s a lot more complexity, and the activities of compromised devices (especially the synchronous activities of botnets) can consume significant bandwidth and state within a mobile network. This can impact service performance and availability. I’m already getting questions on this when I talk to ISPs who have both wireline and mobile networks, as there’s growing concern about how to extend their threat detection and management capabilities beyond their wireline environments to within mobile networks.
This concern is driven by the challenge of getting the right level of visibility into mobile consumer traffic. User traffic within mobile networks runs over GPRS tunneling protocol user-plane (GTP-u) tunnels, making inspection more difficult. And, with traffic levels growing quickly, the inline deep packet inspection (DPI) solutions that have been relied upon by some mobile operators for threat detection are becoming less appealing, due to the near-constant need to re-engineer them as additional network capacity is deployed.
The need for scalable, passively acquired, high-fidelity visibility into the user traffic within GTP-u tunnels is clear. This gives full north-south and east-west visibility into the traffic on a mobile network, allowing mobile operators to quickly identify anomalies that might lead to service problems, as well as giving them a granular picture of network activity that can be used for over-the-top (OTT) content delivery optimization, capacity planning, and many other use cases.
This sounds like a way forward to detect threats, but what should operators do about a threat when they find it? Again, we have to consider the differences between wireline and mobile networks. In a wireline environment, we are all used to detecting threats by internet protocol (IP) address, and then using a variety of mechanisms such as BGP flowspec, BGP redirection, and access control lists (ACLs), which all perform actions based (at their heart) on an IP match. In the mobile world, given the use of tunnels and the lack of BGP, things are different.
There are multiple options from leveraging firewalls to deploying inline devices that can look inside GTP-u tunnels, but in my conversations with mobile operators, the clear winner is network policy. Mobile networks have a much more evolved policy control than wireline, and this policy control can be used to limit, re-direct, and manipulate user traffic. Most mobile operators already use policy control for things like quota violations, so there is nothing to fear here—but there is a downside.
To work with policy control, we need to know more about the traffic we want to manage. Specifically, we need to know about the identity of the device generating it—not just the IP address it has at the moment for a specific service. This means we need to map the IP address we see in the mobile user plane traffic to something like the IMSI (that long number on your SIM card). This requires correlating the user traffic with the control plane procedures going on within mobile networks—this isn’t trivial, but some solutions can do this.
When done successfully, you don’t just get the IP and IMSI, you also get information on the service a device is accessing, the device type, and potentially the cell it is connected to. All of these are useful when we consider a threat, as they can tell us more about where the impact might be or whether there are specific types of devices that are more frequently compromised. And, we can then use our policy control to protect service availability and performance.
One thing is certain, mobile services will become ever more prevalent over the next decade. This will be mostly transparent from a user perspective, or at least things will only get better. As for mobile operators, they will increasingly see the same problems that their wireline colleagues have dealt with for the last decade—and they will need to adopt new operational processes and technologies to manage those problems.
Darren Anstee is CTO for Security at NETSCOUT, which helps assure digital business services against disruptions.