Cloud-Native Infrastructure Automation – The key to 5G success
5G has proven to be a game-changer for several businesses. Given the advancements in O-RAN contributed by cloud-native design & 5G Core, telecommunication vendors, Communication Service Providers (CSPs), and enterprises are trying to deliver an extraordinary customer experience by leveraging 5G. This also presents a massive opportunity for service providers to simplify and enhance customer experience, fortify existing revenue streams, and tap into new markets.
But, this comes with its challenges. A complex set of use cases and an even more complex infrastructure with cloud-native technologies such as Kubernetes is something that CSPs haven’t dealt with before. And that automatically translates into heavy upfront investment in these areas. Consequently, there is a need to control the operational expenditure while offering network continuity and the ultimate customer experience.
Today’s CSPs are already pivoting to a dynamic and heterogeneous model to cater to the needs of highly demanding customers and maintain SLAs. The operations will only get more complex with the constant changes in 5G network parameters, configurations, hardware & firmware upgrades, and new feature rollouts to meet customer expectations. More critically, it is impossible to meet the latency and availability requirements of 5G applications through manual and misaligned processes.
To keep up with the flexibility of 5G and unleash its full potential, CSPs need to fully embrace network automation at the infrastructure level with fully autonomous networks as their North star.
Four critical areas that will give CSPs a clear head start in their journey to 5G supremacy include :
1. Accelerate service delivery & standardize operational procedures
Network slicing will be a key business driver for CSPs. As they gear up to deliver the first use cases of deploying DU, CU, and 5G Core, they need to ensure end-to-end automation readiness to elevate the customer experience. They must work around the changing nature of 5G requirements through service lifecycle management, including instantiation of service chains, upgrades, horizontal and vertical scaling, and termination of applications spanning multiple data centers.
But there is still a missing piece to this puzzle. The operational aspects are often missed and backfilled manually as an afterthought. What completes this, therefore, is an automated process workflow that ties the service lifecycle management to a set of validations and other ecosystem services to ensure a zero-touch network slice/infrastructure management system.
Bringing up the network infrastructure is an easier problem to solve. While Kubernetes has now become the standard for 5G deployments, it will not become the silver bullet unless managed right. Visibility, deployment, resource management, and quality of service form the foundation of a healthy production cluster.
Maintaining the infrastructure to meet the customer requirements can be daunting given the stringent SLA requirements. With enterprises as end-customers, the SLAs could be in the form of mandatory configurations to conform to PCI or HIPPA standards or compliance to the network slice security through a set of KPIs. But how does one ensure the required configurations are always in place? How does compliance go beyond an audit that is just a snapshot in time? A flexible framework for policy definition, scheduled audit, notification & remediation to maintain infrastructure policies for diverse customer segments will ensure compliance with a standard set of configurations.
While this satisfies the configuration requirements, CSPs must build operational compliance on a closed-loop system. It demands an integrated approach by combining automation & the observability framework. The key here is the data collection that varies from bare-metal metrics, the management controller metrics, Kubernetes cluster metrics, and vendor-specific metrics. Constant comparison of these metrics against thresholds offers actionable data to the automation framework. When combined with log analytics, event monitoring and notifications, the observability framework becomes a potent tool for the CloudOps team. This approach helps CSPs to ensure 24×7 availability of the infrastructure and applications.
2. Establish an end-to-end automation lifecycle
Traditionally, CSPs have dealt with legacy networks in 4G and 3G, and the expectations never pushed them into an agile service and software delivery model. They have experienced long product and service development cycles, slow service rollout, manual network monitoring & resolution. 5G differs from legacy systems management in that operational & service agility will be the de facto standard. Infrastructure lifecycle management plays an important role. Gone are the days of proprietary hardware, and today’s infrastructure is based on x86 hardware that can be managed easily due to its rich set of APIs. This agility on the server infrastructure and cloud-native technologies also means a constant set of feature releases and changes to the network. CSPs need to embrace Infrastructure-as-a-Code and introduce CI/CD to accelerate service delivery, validation, and deployment to be the latest and greatest at all times right from the bare-metal.
3. Choose a scalable software stack
Globally, 5G is set to cover more than 20% of the total connections by 2025 as per the data presented by GSMA.. It is evident that the 5G network will dictate the scale. 3GPP adopted a cloud strategy for the 5G infrastructure, keeping this in mind and the dynamic use cases. So how can a legacy & non-scalable automation solution be your partner for 5G? CSPs need to choose automation that can scale with the network to align with their cloud strategy. A modern stack on microservices backed by Open APIs, with multi-cloud support, unwavering security, centralized logging for correlation, with the in-depth backing of infrastructure automation & slice management use cases will step up the game for them.
As CSPs globally roll out and prepare their network for 5G, we witness a wider adoption of programmable networks. Telecommunication & infrastructure vendors have made faster strides by packaging these into their platforms in the last couple of years.
The stage is ready, the use cases are available, but how can service providers take advantage of them?
What the CSPs need today is a single-pane-view of serving multi-vendor networks. The varied data sets discussed above will lay the foundation for the CSPs’ AIOps journey. As the network and cloud infrastructure gets modernized, it is time to look beyond the legacy automation tricks and adopt a holistic solution that will empower the CSPs to plan, deploy and manage networks for their future 5G success.