Telco network healing and auto-scaling with Open Source MANO TWELVE

The second Open Source MANO (OSM) release with Long Term Support (LTS) has landed. Open Source MANO Release TWELVE offers two years of support with security patches. The release has outstanding features to offer to both VNF vendors and system integrators residing in the MANO (Management and Network Orchestration) ecosystem. ETSI OSM (Open Source MANO) can be integrated with multiple cloud platforms and virtual infrastructure managers (VIM). Service providers and operators leverage the OSM platform to run services either on virtual machines (VMs) or containerised frameworks, i.e Kubernetes. This release offers the much-awaited features of auto-scaling and auto-healing for telco networks.

The model-driven approach in OSM makes Network Functions (NFs) reusable and portable. Reusability offers the following operational benefits: 

• Layer abstraction reduces complexity. 
• Errors are minimised with fewer changes required in packages. 
• Optimised deployment times. 
• Minimal human efforts are required.
• Multi-site reliable deployments.
• Ideal for CI/CD pipelines to execute testing. 
• Advanced redundancy mechanisms to make services highly available.

OSM release TWELVE’s features are more aligned with telco industry use cases and demands. Key additions include:

• Self-healing of network functions: Release TWELVE possesses novel features for automated network function healing. The intent for VNFs is specified via VNF descriptors including healing capabilities. For example, if an instance has crashed due to failure, human intervention is no longer needed. OSM will automatically recover the instance from failure with its auto-healing functionality. Manual healing is also enabled through the latest API  developed under GS NFV-SOL005 ETSI specifications.
• Support for ETSI SOL003 specification: OSM integrated a new API developed under the ETSI SOL003 specification. This API enables communication between the VNF manager and the orchestrator. It enables third-party solutions to manage NF lifecycles easily, including operations like creating, managing, updating, and deleting VNFs.
• Update of network function  Day-2 operations: the release includes support for upgrading the existing, running instances of Network Services (NS).
• Support for anti-affinity: VNFs now run in high availability using anti-affinity groups. The group is defined in description intent ensuring VDUs for NF are deployed on multiple hosts enabling redundancy. This feature is supported in OpenStack only.
• Extensions for CNF/K8s support:  this release matures the Day-2 primitives of CNFs. OSM is empowered to discover services managed by proxy charms using its Kubernetes deployment units (KDUs). It is achieved through intent definitions for VNFs or CNFs descriptors. Restricted K8s cluster initialisation and registration for Helm-based KDU or Juju-based KDU is also enabled.
• Security:  Improved security measures are now supported including forced password management on the first login and password expiry after the defined time limit. 
• Support for NFVI level metrics that has been monitored from Prometheus as part of OpenStack telemetry. 
• Support for vertical scaling of OpenStack VM instances. 
• Support for OpenStack VM instance migrations. 
• Support for improved alarm notification channels.

This webinar for Release TWELVE zooms in on the following few features from the list above:

• Upgrading a running instance of NF to use updated descriptors including policies like auto-scaling and auto-healing. 
• Removal of an existing NF instance from a running NS. 
• Changing operational states for NF and its vDU. 

Live upgrade

Seamless upgrades to running CNF instances give better flexibility to apply Day-2 primitives without having to stop them first. It enables operators to apply new changes. The release TWELVE webinar demonstrated a similar feature by deploying a CNF on the Kubernetes cluster. Day-2 primitives were performed on a running CNF.

Telco network auto-healing 

The ability of a platform to recover automatically from failures is a good use case for operators. In the release webinar, one of the 5G core components, AMF, was deleted to showcase how OSM behaves against sudden failures. Instantly, a new instance gets launched to match desired intent in VNF descriptors. The auto-healing feature is specified as intent in descriptors and lets OSM take care of failures with its zero-touch automation.

Telco network auto-scaling

Similarly, the auto-scaling policy is defined in the VNF descriptor to scale-out VNF instances against a threshold. In the webinar, the traffic for UPF instance has been increased to hit the threshold for CPU utilisation and as a result OSM triggered a scaling action for UPF instance to handle the additional traffic. The overall process is instantaneous and smooth and no user has been affected. When the traffic load decreases, the VNFs automatically scale into the actual state. This automatic scaling feature offers optimal resource utilisation for a network service.

Summary

OSM Release TWELVE brings valuable features to OSM by targeting practical use cases. VNF vendors and service providers can leverage new LTS releases and benefit from multi-cloud platforms and live upgrade compatibility. OSM is at the heart of the NFV domain delivering compelling features for innovators to focus on cost optimisation, network automation and enhancements as technology evolves. Release TWELVE delivers features like auto-scaling and auto-healing to telcos for their mission critical and production grade networks. It enables flexibility and agility for networks so VNF/CNF vendors can focus on delivering quality services.