5G and Kubernetes are both hot topics as we enter a new decade. And although each technology offers plenty of benefits on itys own, 5G and Kubernetes promise even more innovation when they converge. That, at least, is what some folks believe, as they promote Kubernetes as the secret ingredient for closing the gap between 5G’s promises and 5G problems.
Kubernetes does indeed have some potential to make 5G actually work well. That said, it’s not necessarily the holy grail of edge computing and telco networking that it is sometimes made out to be. Here’s why.
You don’t have to be a network architect to know about 5G and the revolutionary features it can theoretically deliver. Presumably in a bid to convince customers to buy more expensive 5G-compatible phones, telcos and their partners have eagerly been promoting 5G to the average consumer. They make promises like “5G networks will one day offer peak data rates of up to 10 Gbps” and “5G’s hyper-fast speeds will revolutionize the way we live.”
If you read statements like those carefully, you’ll notice that they’re predicated on theoretical future developments, not what 5G is actually delivering to the typical user today.
That’s because, to date, 5G network implementations haven’t been all that impressive. They turn out to be slower than 4G in many cases, not to mention less reliable. 5G, in other words, has become “a bad joke.”
Can Kubernetes Help Solve 5G Problems?
There are a variety of reasons why 5G has proved disappointing so far. Many of them have to do with the fundamental design of 5G technology and the infrastructure that powers it. 5G has a smaller range, which makes coverage more difficult. Not all devices that boast 5G compatibility are actually supported on all carriers’ 5G networks. 5G infrastructure rollout is limited, which also contributes to poor coverage.
Kubernetes can’t solve those 5G problems. It potentially can, however, address another core challenge hindering 5G adoption: the fact that application deployment strategies on 5G networks are not fast or agile enough to deliver fully on the speed and reliability that 5G can theoretically offer. Today, 5G software development and deployment has a real legacy feel. It falls far short of delivering the scalable, service-oriented experience of modern clouds.
In other words, as ABI Research put it in 2019, 5G networks require a cloud-native development platform to achieve their stated goals. That’s the only way to make 5G services easier for telcos’ partners to consume, build upon and deliver to their own customers.
If you want to build a cloud-native environment today, of course, Kubernetes is the go-to solution. Hence sentiment promoting Kubernetes as the salve that 5G desperately needs.
The Limitations of 5G and Kubernetes
There’s good reason to believe that Kubernetes can address some of the core challenges facing 5G rollouts, at least when it comes to making 5G services easy for software vendors to consume. It’s hard to argue against the idea that, to achieve its full potential, 5G needs to deliver a cloud-native experience in which applications can be containerized and deployed continuously across large-scale infrastructure.
On the other hand, the big challenge here is that transforming legacy carrier environments into Kubernetes clusters will take a long time--far longer than migrating conventional infrastructure to Kubernetes. Not only will telcos and their partners need to find ways to convert legacy workloads that can’t easily be lifted and shifted into containers, but they’ll also have to solve issues like the need for Kubernetes to operate with a plethora of network interfaces, something it’s not designed to do well by default.
Plus, telcos will have to acquire sufficient expertise in Kubernetes--and containerization in general--to oversee their migration to Kubernetes. That’s a tall order, given that Kubernetes engineers remain in short supply.
And even if 5G networks do go cloud-native via Kubernetes, the change won’t solve the other types of 5G problems described above. 5G networks will remain slow and unreliable as long as infrastructure coverage is weak.
The bottom line: Kubernetes has real potential to address some 5G problems. But it won’t address all 5G problems, and it will take a long time--several years, most likely--before Kubernetes is a full-fledged 5G solution.