Jio Platforms goes DIY on autonomous operations with in-house cloud native stack

Who:

Jio Platforms

What:

Developed cloud native autonomous operations stack in-house covering Hybrid Cloud Deployments, Management and Orchestration (MANO), Network Management System (NMS) and Adaptive Troubleshooting, Operations and Management Platform (ATOM)

How:

Solution design is aligned to TM Forum Business Process and Information Framework models and leveraging Open APIs; compliant with 3GPP and ETSI for easy integration

Results:

TCO cost savings estimated to exceed US$1.5 billion over five years; 4G call drops are down to less than 0.1%; 90% reduction in annual maintenance costs and highest customer NPS in Jio’s history.

Many telecoms operators grasp the notion that automation is the future of network operations. For Indian telco Reliance Jio, the future is now.

In fact, operators don’t really have much choice – as user-driven services become the norm, automation is the only realistic means of dynamically managing, orchestrating and coordinating high traffic volumes and complex services across all domains.

The current status quo of network elements (that is, control, management and user data traffic functions) being physically bundled together and provided by the same supplier imposes serious limitations on operational flexibility and increases the likelihood of vendor lock-in. This also slows down deployment of new services, upgrades or modifications to existing services, as these have to be done element by element.

“Future telecom networks will not only have to support millions of subscribers, but also need to support connected intelligence by enabling millions of devices,” says Aayush Bhatnagar, Senior Vice President at Jio Platforms, the holding company of Reliance Jio. “There will be continuous need for on-demand scaling of the telecom network to meet the exponentially growing data and signalling requirements. Hence the operational procedures of modern networks must be automated.”

There are plenty of solutions out there to help operators achieve that level of automation. Jio Platforms decided to take the DIY route by developing its own autonomous operations stack. The solution – which is based on several TM Forum assets and is compliant with 3GPP and ETSI standards for easy integration – enables end-to-end automation, lifecycle management and orchestration of network functions and cloud infrastructure. All the solutions within the stack have already been deployed in Reliance Jio’s network, and the results indicate it was worth every R&D rupee.

Building the stack

The Autonomous Operations stack deployed in Jio’s network comprises four solutions that break down as follows:

• Jio Automated Cloud Installer (ACI): this fully automates cloud deployment on production-grade environments based on Docker and Openstack. The solution enables Jio to quickly deploy hybrid cloud environments in a way that makes them simple to operate, upgrade and scale.
• Jio Converged Management and Orchestration Platform (C-MANO) for network function lifecycle management: this is a microservices-based, distributed, scalable and extensible framework that efficiently scales resources on demand, supports services in a distributed manner, proactively monitors distributed resources and takes intelligent action to manage and orchestrate them.
• Jio Network Management System (NMS): this platform is fully Integrated with AI and ML capabilities to monitor and manage FCAPs for all 5G/4G network elements, providing comprehensive dashboards for easy visualization.
• Jio Adaptive Troubleshooting, Operations and Management (ATOM): this platform leverages ML to detect anomalous network patterns and create reports and alerts based on these patterns before they have a chance to affect operations. It also enables automated SLA management capabilities in the workflow engine, orchestrating operational tasks between the systems.

All of these solutions are aligned to TM Forum Business Process Framework (eTOM) and TM Forum Information Framework (SID) models.

Aayush Bhatnagar, Senior Vice President at Jio Platforms, says the solutions follow frameworks offered by Business Process Framework Aggregate Business Entities (ABEs) defined by TM Forum: Configure & Activate Resource, Track & Manage Resource Provisioning, Report Resource Provisioning, Close Resource Order and Recover Resource. The design is also aligned to the Business Process and Information Framework ABEs: Resource Configuration, Resource Order, Resource Specification, Account, Business Interaction and Users & Roles.

Additionally, all the solutions are cloud native solutions designed on the top of TM Forum Open APIs such as the Product Inventory Management API, Service Activation and Configuration API, Service Inventory Management API and Resource Inventory Management API.

These are a strong starting point for designing microservices because they define a consistent and simple-to-use resource model. By re-using TM Forum assets, Jio Platforms have already seen significant savings in design efforts for these systems.

Life after automation

The above four solutions were developed, tested and deployed in-house with minor and major releases, in sync with the proposed evolutionary path specified in ETSI standards. Jio took a modular approach, with all the microservices divided into layers/categories.

Currently, Jio ATOM is providing insights on over 200 billion data points per day, while 500 cloud native pods have been deployed via Jio MANO for Jio’s 5G core network. Jio ACI has enabled automated deployment of the 5G core network on 44 data centers and over 100 edge sites, while Jio NMS enables all 5G nodes to be pre-integrated, with over 80 4G nodes integrated in the production network, which has resulted in a 90% reduction in annual maintenance costs.

Jio Platforms lists several benefits that the operator is now enjoying with its autonomous operations capabilities, from cost reduction and increased productivity to better system performance and reliability. Automated configuration of data and changes in the network have resulted in improved service quality and performance. Network monitoring and optimization in terms of usage and availability of resources enables operations to identify the root cause of issues.Automation also enables model-driven repeatable service deployment through orchestration, as well as policy-based workload placement for better performance and cost optimization for specific service needs.

Jio Platforms says the autonomous operations stack also future-proofs Jio’s service deployment ambitions by enabling rapid deployment, provisioning and orchestration of network functions and services across networks – which of course creates new revenue opportunities and speeds up “time-to-cash.”

Jio Platforms also points to a recent speed test report from Indian telecoms regulator TRAI, in which Jio topped the 4G average download speed chart at 20.7 Mbps (in second-place, Vodafone Idea came in at 6.3 Mbps), clearly illustrating the competitive difference automation has made in download speeds.

“In addition to that, having their own 4G IMS solution plus ATOM automation has lowered Jio’s call drops to less than 0.1% for the first time, and resulted in the highest NPS customer scores in Jio’s history,” Bhatnagar says.