Skip to content 
Next-generation 5G networks will have to be deployed 100x faster than the traditional telecom solutions, not just in bandwidth but also with customer responsiveness. This is akin to asking a weight-lifting athlete to run a 100-meter sprint. Such a massive challenge can only be solved with technological advances.
Fortunately, Cloud-Service Providers (CSPs) such as AWS, Azure, and Google Cloud Platform have proven that cloud-native, containerized applications can make this a reality.
Anuta Networks is at the forefront of network automation innovation. We’ve been delivering cloud-native network automation for the last four years. This blog will examine the merits of a cloud-native network automation and how it can revolutionize telco networks.
Requirements for Next-Generation Network Automation
From our interactions with global Tier-1 telecommunications providers, it is apparent that next-generation network automation must satisfy the following requirements:
Elasticity – 5G applications will have unpredictable demands due to variations in adoption. A telco network must be adaptive to fluctuating demands.
Rapid Innovation (CI/CD) – 5G will unlock productivity for enterprises and consumers at an unprecedented level. Telco networks, therefore, must be hyper-responsive to support varying complex requirements spanning security, latency, scale, and performance. Telcos can no longer afford to onboard services in weeks in a waterfall model. They must deliver in minutes or risk losing subscribers and market share.
Geo-Redundancy – The pandemic has mandated innovation and a more scalable deployment model across the globe. The new normal is that the future for enterprises will entail the support for thousands of branches or home offices. Hence, Geo-Redundancy is imperative for telco networks.
Security and Control – Network automation represents the keys to the network kingdom. If it is hacked, it’s going to be game over for network operators. Thus, features such as multi-tenancy and role-based access are crucial for security. Multi-tenancy also improves economies of scale, given resources are shared among tenants.
High Availability – Networks are becoming mission-critical. Whether it’s your kid’s school or your company presentation, network outage is very costly. Hence, high availability is crucial for maximum uptime and customer satisfaction.
Visibility and Analytics – Next-generation applications will require complex interactions among hundreds of microservices. Hence, visibility, analytics, metering, and billing need to be integrated into the fabric of the network.
Integrated Platform – Traditionally, telcos rely on best-of-breed tools – spanning provisioning, monitoring, and incident resolution. An integrated solution is mandatory to simplify network operations, provide visibility and generate analytics.
Unified Interface – The smartphone revolution established that user experience trumps the depth of functionality. UX is critical, not just for end-users but also for network operators and architects. Thus, a unified interface will accelerate troubleshooting and optimize the time of skilled network architects vs. a collection of disparate dashboards.
Scale and performance – 5G networks require low-latency and high throughput provisioning and monitoring.
Pitfalls of Legacy, Monolithic Solutions
There is a prevalent trend known as “Cloud-Washing” in the network automation industry. Many legacy monolithic software platforms packaged in a virtual machine (VM) and deployed in the public cloud are deemed “Cloud-Ready Software.” This approach is not only expensive but also extremely dangerous with security and availability ramifications.
Monolithic – Monolithic architectures create complexity. For example, if one module such as device discovery fails, the entire software platform needs to be restarted. Similarly, if you want to upgrade one function, you must upgrade the whole software platform. The bottom line, monolithic solutions do not scale and are cumbersome to manage.
Security Lapses – Monolithic software is inherently insecure due to complex dependencies among multiple assembled functions without any security controls. Many cross-site scripting attacks leverage these vulnerabilities.
Downtime – Most legacy monolithic solutions usually possess an active-standby mode that results in downtime and inefficient computing and license resources usage.
Siloed Solutions – Legacy monolithic solutions follow a siloed approach where operators need 3 or 4 products that don’t align with each other in terms of capacity, capability, or performance. There is also a steep learning curve for multiple products.
Swivel Chair Approach – Multiple products result in distributing crucial network information into islands of automation. Valuable insights are lost, and the meantime to repair will be excessively high.
Performance Bottlenecks – Legacy monolithic software platforms suffer from performance bottlenecks. To keep up with the demand, operators end up with multiple copies of software. This approach is inefficient, expensive, and cannot scale to meet the 5G demands.
Introduction to Cloud-Native Architecture
Microservices are at the core of any cloud-native architecture. Each microservice is dedicated to a differentiated function with its own resource model, APIs, and execution stack. The microservices are typically loosely coupled, and they often interact via APIs. Each of the microservices can be upgraded or scaled independently without needing to affect the whole application. Further, microservices are packaged into containers (such as Docker) that can move from one cloud to another without any modifications. The container orchestration systems such as Kubernetes can also monitor each microservice’s health, latency, and performance.
Cloud-Native Network Automation with Anuta Networks ATOM
Anuta Networks ATOM comprises many Docker containers, each providing a specific function such as device discovery, telemetry collection, network analytics, service assurance, reporting, and more. Kubernetes also manages the deployment of multiple containers.
Technical Advantages of Cloud-Native Network Automation
Auto-Scale – Kubernetes cluster management software automatically scales each Docker container depending on each microservice’s demand and key health metrics. Auto-scale relieves the administrator burden tied to capacity planning. Coupled with elastic compute and storage, the microservices-based ATOM platform can automatically expand its capacity to match the demand for network services. For instance, if the alerting microservice is overloaded and the response time has deteriorated, Kubernetes will instantiate additional microservices to improve the response time.
Rolling Upgrades – Microservices architecture enables administrators to upgrade individual microservices one at a time. This practice will avoid downtime for mission-critical networks.
Multi-Cloud Support – The Docker containers can be deployed on-premises and in public clouds such as AWS, Azure, and Google.
Flexible Deployment Options – Administrators can package the ATOM platform to include only the necessary microservices per customer requirements or budget. The overall software footprint will be smaller, and risks are minimized.
Pluggability – ATOM’s microservices architecture enables fast and easy integration with other enterprise software such as performance monitoring, VNF management, and Machine Learning/AI systems that can be deployed as Docker containers.
Business Benefits of Cloud-Native Architecture
The above technical advantages result in many business benefits. Among these:
Security and Compliance – A multi-tenant architecture with dedicated namespaces for each tenant is secure and compliant with industry regulations. Multi-tenancy delivers economies of scale.
Cost Reduction – A modular architecture with efficient resource utilization reduces costs.
Feature Velocity – With CI/CD practices, DevOps teams can publish new features and services with speed and improved agility.
Horizontal Scale: Cloud-native architectures help scale network infrastructure automatically with the help of microservices.
Furthermore, a cloud-native architecture delivers hitless upgrades, as well as resilient and high availability. Cloud-native architectures also facilitate simple integration with ML & AIOps to deliver valuable insights and positive business outcomes.
Cloud-Native Network Automation Deployment
A significant advantage of a cloud-native architecture is its deployment flexibility and support for multi-site distributed deployments. In this sample Anuta ATOM deployment, the ATOM agents (a stateless Docker container) are deployed in the local AWS zone. The positive aspect of this approach is that the administrator must manage only one server instance in the cloud through a single pane of glass.
All the data is also consolidated into a clustered database that can scale horizontally. The result is that applications, CI/CD tools, and administrators must only deal with one ATOM instance.
ATOM has been built with multi-tenancy from Day-1. For example, the same ATOM instance can be completely isolated for network/organization domains, such as enterprise and retail customers.
ATOM’s cloud-native architecture has also been validated to deliver high performance on a horizontal scale. For example, one instance of ATOM automates 2,000 workflows, each containing 50 pre-and-post checks within 3 hours. Furthermore, ATOM can provision an astounding 200,000 VPN services per hour, with 20 seconds per service latency.
Summary
5G deployments require next-generation network automation to deliver high throughput, low-latency provisioning, elastic scale, geo-redundancy, and high availability. Cloud-Washing and legacy monolithic architectures will undeniably result in expensive mistakes and delays.
Clearly, cloud-native architectures deliver all the technical advantages to provide superior performance and an enhanced user experience while achieving economies of scale and rapid innovation.
To learn more about ATOM Cloud, visit here.
