One hand to shake - Ericsson Cloud RAN verified on cloud native infrastructure solution

At Ericsson we understand that the stability of telco cloud platforms has been a challenge in the past, but it has now reached a high maturity level, and we have gained invaluable experience from our many live cloud-based core networks . A major contributor to this is our system-verified methodology in building cloud infrastructure solutions which we describe below.

Through our experience of working with cloud solutions, we have developed a philosophy for how cloud infrastructure should be designed and deployed. It addresses CSPs’ challenges by delivering system-verified cloud infrastructure solutions, which work out of the box.

Key learnings

• Pre-integrated configurations to reduce complexity, lower costs, and faster time-to-market for new solutions are crucial.
• Ensure that all layers in the network, including applications, infrastructure, management, and orchestration, work perfectly by using a system-verified approach.
• Enhancing automation in all parts of the network to address complexity, remove manual mistakes and increase efficiency, is a top priority.

The journey towards Cloud RAN

Ericsson Cloud RAN is the virtualization of RAN functions that run on commercial off-the-shelf (COTS) hardware, and it brings RAN programmability to add new functions and value. Management, orchestration and automation ensure smooth and efficient life cycle management of RAN services, applications and cloud infrastructure, which we have explained in our blog: ‘What is Cloud RAN and its four key components?’

Cloud RAN opens new opportunities for CSPs, giving them the ability to scale when needed and the possibility to collaborate with partners using standardized interfaces, which allows them to explore new propositions.

Ericsson Pre-Integrated Cloud RAN Solution

With openness comes complexity, and adopting Cloud RAN often means making significant changes to the existing models of CSPs for RAN development, procurements, deployment, and operations. Ericsson understands these complexities and offers pre-integrated options for Cloud RAN that lower the barriers to Cloud RAN deployment - you can read more in our blog Pre-integration – lowering the barrier for deployment of Cloud RAN.

Pre-integration helps CSPs to get a head start to their Cloud RAN journey. One option is the full-stack system-verified Ericsson Cloud RAN solution based on Ericsson solutions and products. This brings simplicity, speed, and agility into the CSP’s deployment of Cloud RAN.

Pre-packaged

Building a virtualized RAN over a cloud infrastructure is similar to putting together children’s building blocks. These blocks can be used to create wonderful models, with limitless customization possibilities in different layers. When you walk into the toy store you will often see the blocks pre-packed in different sets, and there are also “pick-and-build” elements that offer you the flexibility to choose the ones you wish to use for your own creations.

The pre-packaged sets of these blocks are pre-integrated and well-tested, with clearly documented step-by-step instructions for how to build and use them - if you follow the instructions exactly as described, the outcome will be perfect, and it will turn out the same as shown on the packaging.

With the “pick-and-build" variety, the added flexibility needed to build your own design comes with time and cost. You are basically responsible for all the pieces by yourself and it's your own creativity, knowledge, and execution that creates the whole solution stack. You prepare the idea, choose the blocks, and build everything as you wish from scratch; if you get stuck with the wrong piece, it’s up to you to fix and improve, which might involve going back and forth to the toy store. This approach is well-suited for experts who have plenty of experience of building the pre-packaged sets and who have the money and passion to build something personalized.

A telecom optimized system-verified solution

Building a cloud infrastructure and optimizing it for telco applications has always proven to be complex, for both vendors and CSPs. A full-stack system-verified solution is based on a reference configuration – a pre-packaged solution - including Network functions, Management and Orchestration and Cloud Infrastructure that are well integrated and tested to make up the full-stack solution. A system-verified solution is created through extensive system design, verification, documentation and life-cycle management of selected software and hardware components. This approach towards cloud transformation of network functions has proven not only to be efficient, but it also enables faster time to market for the solutions.

The journey begins with defining a reference architecture involving product selection to realize a flexible and scalable architecture that can address different use cases. To be well-integrated with the existing network components and operations systems, the solution architecture needs to comply with open standards like ETSI and 3GPP specifications.

The next step in the process is realizing the reference architecture using the selected hardware and software products by integrating and verifying the solution across functional and non-functional areas. This activity is the most critical part of the process, involving many R&D integration engineers and solution architects, as well as different testing lab environments. The integration and verification effort needs to be redone for every solution release, to support the life-cycle management process of the different components in the stack.

The integration and verification process provides valuable insights into system behavior, especially application performance when the system is subjected to close to real-time network environments. These insights help to identify and drive the required optimization in different layers, including fine tuning of compute server BIOS and the operating system kernel, as well as optimization efforts in application.

Another important step in system verification is the application orchestration and life-cycle management validation, where application-specific life-cycle management (LCM) operations like onboarding, upgrade, scaling (in and out ) and maintenance are thoroughly tested over a given Cloud infrastructure SW baseline .This validation ensures that the compatibility between application and underlying cloud execution environment is in place, and also that there is no system performance or characteristics degradation induced due to changes in the SW baseline.

As part of the pre-integration and solution verification process, automation is crucial to rolling out features and new software at a much faster pace, not only in the internal development environment but also in customer networks. With a system-verified stack, implementing automation across the solution becomes much easier as we understand the granularity of automation needed in different unique layers in the stack.

What is the full-stack system-verified Ericsson Cloud RAN solution?

Ericsson’s full-stack system-verified offering makes use of our field-proven Cloud Native Infrastructure Solution (CNIS) (read more). In this solution, Cloud RAN applications and CNIS are fully integrated, verified and optimized for faster time to market.

CNIS makes use of the following Ericsson products:

Ericsson Software Defined Infrastructure (SDI) is a local SW entity to manage hardware resources at the Cloud RAN site. It manages generations of hardware resources (compute, networking, and storage) through a common inventory and API. SDI provides HW resource management and interfaces with chosen 3PP compute and networking products as part of the CNIS offering. SDI also enables life cycle management of local site-specific Cloud RAN HW resources, including “day zero” deployment – that is, initial boot procedures, HW authentication and capturing of inventory and “day ‘n’”-related activities like upgrades of firmware.

Ericsson Cloud Container Distribution (CCD) is a system-verified Bare Metal Kubernetes CaaS that is pre-integrated with Ericsson Cloud RAN applications. CCD is based on Cloud Native Computing Foundation (CNCF)-certified CaaS with ‘no forking’ and ‘upstream first’ design principles. It reuses open and common APIs to maintain interoperability with 3PP hardware and software components.

Ericsson Operation Manager Cloud (OMC) infrastructure is a regionally centralized resource orchestration and management system that provides a northbound API which aggregates resource orchestration of highly distributed cloud RAN infrastructure at various RAN hub and cell sites. This northbound API provides the ability to integrate with Service Management and Orchestration (SMO) systems, such as the Ericsson Intelligent Automation Platform.

OMCs using southbound APIs provide central aggregation of distributed hardware resources by interfacing with SDI components at distributed cloud RAN locations. Similarly, OMC will provide a central cluster management capability by interfacing with CCD at the distributed site locations. This setup helps aggregate all fault, performance, and security-related metrics, which help visualize the topology, connectivity, and state of the deployed cloud RAN resources.

OMC provides a common infrastructure integration point, with a common set of tools and templates to achieve automated workflows, thereby contributing to operational efficiency.

The benefits of a full-stack Ericsson Cloud RAN offering

With a full-stack system-verified solution, CSPs will be given a head start in their Cloud RAN deployment by reducing complexity and offering faster time to market. The intent of the offering is to provide a fully pre-integrated solution, meaning it is optimized for Cloud RAN for performance and also includes dimensioning variants for various kinds of deployments. It’s important to note that system-verified solution approach lays a strong foundation for automation and operational benefits that CSPs expects from Cloud Transformation.

As part of the system-verified solution, we include a complete life cycle management, covering all hardware and software components, as well as all aspects of infrastructure and application orchestration and management. This approach ensures that Ericsson, as a single vendor, has responsibility for the end-to-end solution, including fault isolation, reproduction, and correction. Multi-vendor and multi-layer cloud infrastructure stacks always come with the question of who takes the end-to-end responsibility - who invests in the competence needed to support the solution and secure stability and performance?

A system-verified cloud stack provides CSPs with a single point of accountability for solution support tasks, including ecosystem fault isolation, issue and emergency recovery services, root cause analysis, management of the integrity of customers’ solutions and vendor coordination.

Added to that is the comprehensive end-to-end security for the entire solution, as well as a common set of unified tools across applications and infrastructures and aligned roadmaps and coordinated release management between application and infrastructure.

Our conclusion is that we see different approaches to build and realize a cloud infrastructure for Cloud RAN. We believe that a full stack offering comes with clear values for CSPs that wish to be like that child in a toy store who wants a pre-packaged option with a predictable outcome, reliable time to market and ease of mind.

Read more

Ericsson Cloud RAN vision, insights, and ecosystem partnerships 

Explore Cloud Native