“If virtualization is the disaggregation of hardware, then cloud-native is the disaggregation of software.”
As we stand on the cusp of 5G’s arrival next year (according to most analysts), it’s surprising to me how many people in the industry still don’t agree on what 5G will look like when it does arrive. You’ll hear the usual terms in a conversation about 5G —virtualization, cloud-native, the Internet of Things—but little consensus on what constitutes a true 5G network. Is virtualization enough or does a 5G network have to be cloud-native too? Is a cloud-native architecture simply one that features containerized software or is it something more?
What is Cloud-Native Architecture? Here’s a Simple Definition:
Part of the confusion around cloud-native architecture stems from our industry’s affection for making the simple complex. For example, the industry definition for cloud-native invokes a whole laundry list of requirements—containerization, microservices, network slicing, etc.—that many in the industry don’t fully understand. It was during one such conversation, at a recent panel session on 5G networks, where I finally discovered a simple way of defining cloud-native architectures. “If virtualization is the disaggregation of hardware,” I said, “then cloud-native is the disaggregation of software.”
Disaggregation & Cloud Native
For me, that’s the whole crux of the 5G discussion right there. Virtualization is a mandate for 5G, true, but there’s a higher purpose for 5G than squeezing more cost out of the network: a mandate to make money. This requires an innovation platform built on cloud-native architecture principles that can enable new network slices and fail fast modes of operation. And disaggregating software is the key to creating and managing such a platform and driving new revenue opportunities of the 5G future. Achieving the right level of disaggregation is the true art within the science of Cloud Native.
To be clear, cloud-native is more than disaggregated software. It involves, among other things, containerization, dynamic management, an externalized state, a service mesh, and, perhaps most importantly, a microservices-based architecture. Unfortunately, many network equipment vendors are clouding the perception of cloud-native by limiting its scope to a few checked-off boxes. They build on a foundation of hardware virtualization by re-packaging their existing software in containers, provide some dynamic management tools, and announce that their solution is cloud-native architecture. Not surprisingly, their 5G conversations also stop short of promising more revenue. Instead, their focus is on operational and capex savings. Let’s be very clear: If your 5G roadmap ends at network savings, you’ll be running out of road a lot sooner than you think.
Ultimately, 5G is defined by how you define success. If success is saving money and watching other people innovate, then stick to virtualization. But if you define success by shaping the future and embracing the opportunity, cloud-native is a must. There are a lot of roadmaps out there, but in the race to revenue, the path for 5G is clear: the winners will be those who figure out how to drive revenue through new services, faster than the competition.