Tech Mahindra has launched an AI-assisted network operations platform in partnership with AWS that it says has been in “stealth mode” to date, and is now ready to be adopted.
It said that the platform has achieved 50% productivity increase for NOC staff, reduced field visits by 15% and cut mean time to repair by 30%.
It told TMN these stats were “based on our experience of managing and operating networks on the customer’s behalf and the improvements measured on trial deployments with Tech Mahindra’s network customers while the platform was in stealth mode.”
The Autonomous Networks Operations Platform (ANOP) has a three layer architecture: an infrastructure layer, the ANOP platform itself, and then a smart operations layer for which it is developing hundreds of use cases.
Tech Mahindra said it synthesises its own telecoms specific knowledge sets into classical and Generative AI-based ANOP.ai base models. Using its managed data pipeline, these base models are enriched and tuned for a CSP, compliant with data residency and sovereignty requirements provided by AWS AI/ML toolsets.
The AWS Infrastructure layer is combined with AWS-managed AI infrastructure that includes Amazon Bedrock to help create custom LLMs for CSPs, the SageMaker suite of AI/ML toolsets, Managed Data Infrastructure that includes DynamoDB, Lake Formation, and EMR as well as functions such as Amazon Event Bridge, Step Functions, and Amazon Q Developer.
The ANOP platform layer introduces AI models augmented with CSP datasets and consists of functional elements that include:
- Federated Edge Intelligence – Managed on-premises or cloud-based AI for data ingestion, automation, and on-premises or cloud-based arbitration
- Smart Data Curation and Organsation – Platform for data mediation, governance, multi-domain observability, and datasets for AI
- AI Model Lifecycle Management – Manage training, learning, tuning, and testing AI models with deployment orchestration
- Proactive Automation – AI-powered descriptive, predictive, and prescriptive workflow automation at the edge/cloud
The “Telco operations layer” delivers operational use cases through functional segments that include:
- A Cognitive Alarm Handler that can use descriptive and predictive insights for alarm correlation, automated incident creation, and routing modules.
- Proactive Automation, which can help execute tasks to automate incident troubleshooting, diagnosis, and resolution
- Smart Advisory functions that provide prescriptive recommendations, including contextual guidance for scripts, incident resolution steps,and adaptive operational dashboards that are specific to Operations roles for NOC and field teams
- AIML-based correlation of service level metrics with resource level metrics.
TechMahindra said it has built more than 150 use cases across the above functional areas and is developing another 100 additional use cases as a roadmap.
The company told TMN, “Our ANOP solution’s biggest advantage is that it is the most comprehensive solution available in the market to manage end-to-end network operations.”
The company added that such a solution could also help telcos as they roll out new technologies such as Cloud and Open RAN.
It told TMN: “Tech Mahindra has been a strong supporter and has delivered several large projects in deploying Open RAN networks. To accelerate deployment and bridge the gap in industry expectations concerning interoperability and integration of different products, Tech Mahindra is actively involved in testing and verifying such solutions in the market.
“In this regard, we have partnered with AWS to test the readiness of its EKS-A platform for Open RAN CU-DU deployment and verification. Our home-grown NetOps.ai solution further aids CSPs in deploying, testing, and operating Open RAN networks.”
Comments
0