The rewards of public/hybrid clouds for core deployments

To realize 5G’s full potential for offering new consumer experiences and helping enterprises shift to Industry 4.0, most CSPs are deploying their 5G core networks on private cloud or bare metal. Others are considering public or hybrid clouds, with a few already beginning their deployments.

Public clouds include hyperscalers such as Amazon Web Services (AWS), Google Cloud Platform (GCP) and Microsoft Azure. Hybrid clouds coordinate workloads across private and public clouds, or sometimes are public clouds deployed on-premises.

Communications Service Providers (CSPs) choosing the public or hybrid cloud route can enjoy substantial business rewards, but risks must be addressed.

Working with hyperscalers brings several advantages. CSPs can transfer cloud complexity to the hyperscaler, which already has the necessary cloud expertise, infrastructure and systems. For example, faster time-to-market and improved operations efficiency are achieved by taking advantage of hyperscaler-trained talent and their existing CI/CD DevOps, automation and analytics capabilities. CSPs can then focus on service differentiation, not the underlying cloud platforms.

Apps and services can be launched immediately with subsequent optimization of traffic and costs across private and public clouds. New customers with unknown demands can be served flexibly, avoiding potentially wasted CAPEX for an under-used private cloud by shifting to the OPEX of a public cloud. Public clouds can also meet unforeseen demand with added capacity delivered during disasters or large traffic events – CSPs avoid the costs of owning idle capacity.

Partnering with hyperscalers also increases the chances of jointly winning deals with large enterprises. And CSPs gain edge cloud uniformity on a global scale to consistently support a multi-national enterprise’s business processes regardless of location.

But what about those risks I mentioned?

As with any infrastructure choice, there are risks. Using hyperscalers’ public clouds for the 5G core brings new risks that need to be planned for. The most prominent risks (and how they can be solved) include:

• The CSP remains connectivity-oriented, missing out on new lines of business. To avoid this possibility CSPs can engage the hyperscaler’s broad ecosystem to catalyze the creation of new services from the 5G Core’s network APIs.
• The CSP gets locked into a hyperscaler’s public cloud. This risk is mitigated by ensuring the 5G core can use any CaaS/PaaS and insisting the hyperscaler complies with the CSP’s requirements for the 5G core’s Cloud-native Network Functions (CNFs).
• The CSP lacks differentiation if the edge cloud is dominated and defined by a few public cloud providers. This is solved by CSPs exploiting their geographical advantages created by their local workforce, cultural know-how, unique facilities and trusted business relationships.
• The CSP finds itself unable to control its 5G Core to meet business demands. This is avoided by applying strict SLAs and penalties to assure hyperscaler compliance.
• Uncertain cost savings. Avoid the trap of thinking that the 5G core’s business case for public/hybrid cloud deployment relies on Total Cost of Ownership (TCO), which nearly always will be more attractive in a private cloud.
• Increased exposure to intrusions and security risks. Security must be designed into systems, processes and platforms. Ongoing, disciplined and expert attention is vital from the CSP, 5G core supplier and hyperscaler public cloud supplier.
• Risk of non-compliance with regulations on data ownership, geographic location and privacy. A CSP can mitigate this risk by thoroughly assessing and implementing regulatory requirements in contracts with hyperscalers.

Seven principles to guide 5G core deployment on hyperscalers’ public clouds

The Nokia 5G core is deployable on any cloud and prepares CSPs for the journey to Network-as-a-Service (NaaS). We achieve this through seven key design principles. Read more in this article by Jitin Bhandari, CTO & VP, Core Networks, Nokia. Briefly stated, the principles are:

1. Cloud-native foundation: apply hyperscaler technology to CSP networks and operations.
2. Put security first in applications: detailed risk assessment and disciplined design.
3. Openness of networks and ecosystems: apply strong, uniform APIs for a programmable and open 5G.
4. Design for lifecycle inter-dependencies: automate lifecycle management across cloud platforms.
5. Build a service intent network: orchestration designed into slicing and agile vertical services.
6. Multi-vendor continuous delivery: enabling ongoing seamless upgrades with CI/CD.
7. Flexibility to use any cloud platform: design for performance and scale of real-time applications.

It all boils down to the need for each CSP to study the opportunities, risks and design principles by running their own business case to analyze the deployment and operation of the 5G Standalone Core on private versus public or hybrid clouds.

Discover more on:

-    Web page: Core Networks on Public and Hybrid Clouds