Revolutionizing Telecom: Navigating the Shift to Cloud-Native OSS With cloud-native OSS emerging, certain challenges need to be addressed to fully leverage its potential. ? Challenge #1: Integration of Gen AI in OSS Deploying AI in OSS can optimize network management, automate processes, and enhance service delivery. However, issues like data standardization, AI model training, and integration with legacy systems can impede its adoption. Solutions: - Data Standardization: Establishing uniform data formats and common data models is crucial. Industry bodies such as TMF can play a role in setting these standards, promoting interoperability and ease of AI integration. - AI Training and Model Development: Invest in robust training frameworks that use accurate, comprehensive datasets to train AI models. Partnerships with academic institutions and technology companies can enhance AI capabilities in OSS. - Legacy System Integration: Utilize middleware and API gateways to connect AI systems with existing OSS infrastructure, allowing for a phased and flexible integration approach. ? Challenge #2: Implementing Cloud-Native OSS Transitioning from traditional OSS to cloud-native architectures involves significant changes in technology infrastructure and operational practices. Solutions: - Adoption of Microservices: Break down traditional monolithic OSS applications into microservices. - Containerization and Orchestration: Use technologies like Kubernetes to manage containers that encapsulate microservices. - CI/CD: Implement CI/CD pipelines to enable frequent updates and consistent deployment of new features and fixes, reducing downtime and accelerating time to market. ? Challenge #3: Aligning with NGOSS and Event-Driven Architecture Implementing this requires an overhaul of traditional OSS to support event-driven, responsive operations. Solutions: - Event-Driven Processing: Adopt event-driven architectures that react in real-time to network events and customer interactions. - State/Event Management: Develop state/event tables as suggested by the NGOSS Contract to manage complex interactions within the OSS. - Service Modelling and Automation: Utilize service orchestration platforms that align with NGOSS principles, automating the lifecycle management of network services and ensuring compliance with established standards. ? Challenge #4: Fostering Ecosystem Collaboration Solutions: - Open APIs and Interfaces: Standardize APIs as per industry guidelines. - Partner Management Platforms: Develop platforms that enable easy integration and management of third-party services and solutions. - Community and Standards Participation: Engage actively in industry forums and standards organizations. Ref: https://lnkd.in/eBnh_n6q #TelcoDigitalSupportSystem #OperationsSupportSystem #NGOSS #GenAIOSS #AIOps #GenOps #CloudNativeOSS #TMF #ODA
There could be a more radical transformation beyond the Digital Boundary, where the OSS is flattened. The customer/service-centric microservices may descend in the form of modernized AIOps or ITSM. The network control plane functions with event-driven state machines or policy controllers can be generated and instantiated in the distributed multi-cloud/ edge infrastructure. The latter was once a vision of ONAP where SDN control and EMS/NMS/MANO FCAPS weren't bound to the NOC anymore. I think the idea is not completely gone with a network-near SMO that can evolve toward multi-domain automation. The NGOSS challenges may be addressed with both top-down and bottom-up perspectives, using multiple knowledge foundations and service-based architectures.
Reflective of pivotal directions for leveraging the scale of generative artificial intelligence, cloud-native functional portability, combined with an end-to-end overlay of autonomic principles to effectively manage complexity, while optimizing total cost of ownership, performance, and operations, through self-adaptive automation.
Great insights on the challenges and solutions in transitioning to cloud-native OSS! Managing data effectively across diverse applications is indeed critical. Consolidating data into a single source of truth would enhance trust and facilitate automation. Exciting times are ahead for telecom!
Integrating GenAI into the design onboarding process for Cloud-Native network functions could significantly enhance automation, addressing the current gap in automation within this phase.
Brilliant presentation Jinsung Choi, challenge and solution style. Interested to see more details of those solutions and examples of operators who tried them.
Very well articulated and elaborated the challenges and it's solutions..thanks Jinsung Choi for sharing...
Founder and CEO @EnterpriseWeb
9 个月Given Telecom's are infrastructure centric there is a tendency to believe Kubernetes is a magic wand, but it hasn't been the case in practice. The rise of AI (and related Telco struggles) reveal that. A lot happens above the cluster. Except for GenAI, this was all possible a decade ago. Addressing challenges 2-4 enables successful AI and GenAI initiatives (challenge 1).