What does cloud-native mean for network infrastructure?
Communication service providers (CSPs) are on a journey from being connectivity-centric businesses to digital service providers, upgrading their networks to bring customers 5G. This evolution began with the core — the heart of the network — which is on a journey of its own. From the evolved packet core of LTE to the adoption of virtualized packet core, its latest transformation sees CSPs moving to a cloud-based core strategy, something that’ll be characteristic of virtualized radio access network environments.
Cloud computing is a model for enabling ubiquitous, on-demand network access to a shared pool of resources — including networks, servers, storage, applications and services — that can be quickly released with minimal effort or provider interaction. Although payment for use of shared resources isn’t a new concept, the embracing of cloud principles by CSPs is a fairly recent development. Adoption of virtualized and programmable telecom infrastructure, like automation and distributed computing, is being driven by growth in data traffic as well as the complexity of 5G services.
Cloud-based networking is also ongoing. Initially, CSPs put their telecom workloads in the cloud as a kind of “data centre” mechanism. This was an important step to take; but such an implementation mostly comprised of transitioning network elements into virtualized network functions.
Such cloud architecture had no standardized procedures to develop and benchmark virtual network functions, which led to a lack of guidelines and no standard protocols or configuration policies for suppliers. Because they consume too many resources and use older operations, these cloud-based networks are unwieldy to deploy, scale, upgrade and maintain — making them unsustainable.
Demands for 5G performance such as ultralow latency, localized reliability and traffic steering means CSPs need more from the cloud. To achieve the agility for rapidly onboarding new apps and services, the cloud networking evolution is shifting to bring cloud-centric benefits to the network edge. The real value of cloud-based networking for CSPs is in cloud-native networks made for the cloud rather than just being placed in it. A cloud-native approach is the most effective way for CSPs to unlock the full potential of 5G and edge applications.
Essentially this means building and running network functions inside the cloud, based on cloud-native principles and containerization technology — a form of virtualization in which apps run in isolation but share an operating system, with everything each application needs encapsulated individually. This approach is more supportive of application innovation and enables faster development, greater flexibility and service agility. Containers-as-a-service, together with orchestration, can be incorporated as elements of a cloud-based network management system.
Containerization highlights the importance of a telecom cloud-native platform that can efficiently run virtualized network functions as microservices — small, focussed and autonomous services that work together as a distributed system. Microservices are a crucial underlying principle of cloud-native networks; in a 5G network context, this means that functions such as signalling or authentication can be split into sub-services and run in a way that’s more suited to a cloud environment. Examples include running over a service-based interface or a service mesh architecture. This enables greater agility and scalability, possibly as part of a multicloud strategy.
To make this transition simpler and faster for CSPs, a cloud-native network platform is essential. One example of a solution provider at the forefront of this advancement is Intel. Its Xeon Scalable platform brings together a suite of capabilities including software disaggregation, Kubernetes — an open-source platform for managing containerized workloads and services — and massive end-to-end automation. This is designed to enable dynamic, intelligent workload placement and microservices deployment through a services mesh combined with middleware, hardware and operating systems. These capabilities can enable networks to be more automated, use artificial intelligence-driven data analysis and make proactive decisions about resource allocation.
In addition, Intel has built on its ecosystem-creating heritage by developing the Network Builders programme, a community of telecom and IT solution providers working together to accelerate cloud-native network adoption. The programme helps CSPs transition network functions into containerized, cloud-native microservices by giving them access to a system of solutions and best practices. CSPs will choose different paths to implementation; this will enable them to create the most effective architecture for their 5G network.
Cloud-native 5G networks also engender something of a mind-set shift for CSPs. They could enable CSPs to move away, at least in part, from an intensive “build it, own it, operate it” model of infrastructure provision to one based on usage and centred on services offered by cloud service providers. As part of this shift it’s important for CSPs to acclimatize to key performance indicators being different in a virtualized network, compared with a traditional hardware-centric model.
Cloud-native networks could also enable CSPs to share the cost of supporting new technologies such as 5G by deploying shared infrastructure models — either directly through network sharing, or indirectly through investment and offering network-as-a-service in the cloud.
Finally, it becomes crucial to form strategic alliances focussed on cloud providers as service partners; these are well-resourced players with infrastructure capable of supporting every component of the service, providing applications and cloud facilities on a global scale.
Cloud-native networks extend cloud’s benefits for CSPs — a few examples include more efficient use of assets, resource aggregation gains, rationalized hardware inventories, greater network scalability, service agility and resilience as well as operational efficiencies through automation. The demands of 5G means that in the cloud isn’t enough; networks must be built for the cloud. It’s time to go native.