Accelerate intelligent operations using AIOps for cloud native networks
Tue, 30 Jan 2024 17:07:35 -0000
|Read Time: 0 minutes
Dell Technologies infrastructure blocks enable telco customers to adopt telco-centric AIOps to improve operations.
Communication service providers (CSPs) are racing towards fully autonomous networks and consider automation and artificial intelligence (AI) adoption in telecom networks to be of great value. According to the latest industry insights report published by TM Forum® (New-generation intelligent operations: the service-centric transformation path), most CSPs aim to achieve Level-3 automation (conditional autonomous networks) and Level-4 automation (highly autonomous networks) by 2025. There is increased interest in accelerating Level-5 automation (AI-driven automation) using AIOps solutions.
However, telco adoption of AI-centric automation is not easy primarily because most CSPs operate a geographically distributed brownfield network and manage a multi-generation fleet of infrastructure and resources. CSPs also operate at different scales which means there is no simple, “cookie-cutter” approach towards AI-driven operations (AIOps).
In addition, CSPs adopt solutions based on clearly defined, standard telecom architectures like ETSI® (European Telecommunications Standards Institute), TM Forum® (Telecom Management Forum), 3GPP® (3rd Generation Partnership Project), and O-RAN (Open RAN alliance). CSPs also source solutions that can interwork and interoperate at a global scale. Finally, CSPs expect these solutions to fully integrate into their brownfield environments.
Hence, there is a requirement to build an outcome-based solution that supports existing operations. At the same time, the solution must enable them to accelerate the adoption of the next era of operations (based on data-driven insights and artificial intelligence).
How to adopt AIOps-based, intelligent operations for networks
CSPs are working alongside many standards bodies (especially the TM Forum®) to accelerate automation towards Level-4 (full-service orchestration and automation) and Level-5(AI-driven automation). However, there still lacks a clear path for applying these principles to large-scale networks. Building the right architecture starts from clear requirements quantification.
The right AIOps solution that is designed for CSPs must align with unique telco-specific requirements on:
- Distributed topology maps. Telco networks are being purpose-built and deployed to deliver critical and differentiated services for many decades. These networks are not like data centers but instead a fleet of resources—such as home networking, fiber, transport, radio, core, cloud, and WAN services. Topology alignment (like A/B plane, Ring) and service resilience are key requirements.
- Multi-vendor and multi-generation. Typically, CSPs operate a brownfield network over multiple generations. Most of these systems have an extended lifetime of 10 to 15 years. So, the solution should not only be future-proof but also cater to the requirements of existing deployed solutions.
- Data models. CSP networks, by nature, are highly protected—with network data existing in many silos. Network operations also follow a hierarchical, process-based delivery that is defined by Network Operations Centers (NOCs). In addition, data knowledge is based on tools and systems that vendors provide.
Given that AIOps systems are already proven and prevalent in the Cloud and IT industries, these systems and solutions must be adapted to meet CSP requirements.
AIOps in networks should strictly align with both telecom standards and network-specific needs—delivering the following capabilities:
- Process alignment. The current operational model of a telco cloud heavily relies on an operational team knowledge base and expertise. So, it is not just about data but also the unique experience of CSP operations—which are important.
- Data access. CSPs follow strict security and privacy requirements where customer data and information cannot be exposed. So, in order to adopt AIOps, data access models must be standardized to ensure AIOps use cases can retrieve data as per approved policies.
- Tuning. Because CSP-deployed networks must operate for extended periods of time, current solutions—which follow strict AI rules—cannot meet their future requirements. Therefore, AIOps systems must be adaptive.
- Scalability. CSPs operate at different scales starting from Tier-1 (many geographies) to Tier-2 (small scale). Therefore, telco-specific AIOps systems should offer a T-shirt sizing approach.
Accelerate the network AIOps journey
Today, many CSPs have already deployed small-scale AIOps solutions. However, most of these solutions are not highly aligned with telco-specific requirements—resulting in many silos that are hard to manage and scale. Further, CSPs must invest heavily in terms of time and cost to do Life Cycle Management (LCM) of these solutions. It all translates to barriers towards cloud native transformation.
Just as telco cloud CSPs have adopted standards like ETSI®, LFN® (Linux Foundation Networking) and ORAN® (Open RAN), there is a requirement to adopt a standard architecture for the Telco Multicloud AI foundation that can smoothly integrate with brownfield networks. Below are the key capabilities of an AI-centric telco platform that can enable AIOps use cases:
- Horizontal AI platform. The telco-centric AI platform should enable a composable platform that consists of:
- AIOps application layer: hosting various AIOps tasks
- Machine Learning (ML) layer: adopting specific ML models suitable for AIOps
- Knowledge layer: integrating the NOC processes and knowledge of CSPs
- Data layer: resolving any data silos in networks
- Physical layer: managing telecom networks using fully decoupled infrastructure automation
- Distributed data ingestion. The telco-centric AI platform should ingest data from fully distributed networks—delivering both reactive (respond after event occurs) and proactive (predictive) use cases on:
- MOP integration: Existing MOP and workflows must be integrated.
- Operational processes integration: Existing NOC processes must be integrated in data pipelines.
- Cloud native MLOps and AIOps capabilities. Telcos must supplement operational in-house knowledge with ML models and find a way to tune and extend it. Different models must be integrated. A systematic integration of knowledge systems with ML models (in a use case-driven approach) is required for success in network operations.
Figure: Reference architecture for AIOps-driven telecom network
How to adopt AIOps operations using telecom infrastructure blocks
Dell Technologies has worked closely with leading cloud partners, including Wind River® and Red Hat®, to bring forward an operationally ready telco cloud platform. This platform is thoroughly tested, validated, and automated to deliver telco AIOps use cases. This platform also accelerates a CSP’s adoption of zero-touch operations while consistently aligning to telecom standards and frameworks.
The Dell Technologies Telecom Multicloud Foundation flexibly transforms network operations towards programmable infrastructure using a consistent tooling and AIOps capabilities approach.
Because the platform supports multiple versions and offers with various partners, CSPs can operate all such foundational infrastructure blocks as one. Through the following key capabilities, our solution can quickly transform operational models and processes and enables agile MLOps (required in a telco environment).
Figure: Solution architecture for AIOps Foundation
To support the unique CSP requirements to adopt AIOps and a cloud operating model, our Telecom Infrastructure Blocks provide the following key capabilities:
- Consistent platform. The first challenge is to deliver a standard and consistent platform that can integrate all layers above—abstracting the complexities of multiple technologies and components from different vendors.
- Cloud native MLOps. AIOps use cases require cloud-type agility towards data and ML. Our current version of infrastructure blocks delivers a ready platform. In future releases, we plan to enable AI enhancements (like Openshift® AI) on top of this platform. This means CSPs can build, program, and manage all their ML models and capabilities in the same way they manage cloud resources.
- Autonomous operations. Adopting data-centric architectures and ML approaches provides CSPs a smooth evolution path from their current automation approaches to an AI-centric automation that is aligned with the telco future mode of operations (FMO).
- Data-driven architecture: The automation architecture is data-driven and distributed, so data can be tapped from edge and regional sites—enabling real-time use cases and data-driven operations.
- Automated fault management: The FMO follows zero-touch and intent-driven networks. Our solution is fully aligned with this vision that enables all cloud platforms to use declarative workflows. The solution also enables all northbound integration towards orchestration and assurance systems.
- Single pane for DevOps and MLOps operations: As CSPs adopt ML/AI frameworks to deliver AIOps use cases, there is an increasing requirement to integrate and operate both DevOps and MLOPs as one. In addition, AIOps platform capabilities must be enabled in telco cloud platforms. Doing so provides a single management and observation platform.
Figure: DevOps and MLOps workflow using AIOps platform capabilities
Dell Technologies developed Telecom Multicloud Foundation and Telecom Infrastructure Blocks to accelerate telco cloud transformation. Our engineered and factory-integrated system delivers a consistent platform. This platform is ready to deliver telco-specific AIOps use cases that are fully aligned with telecom architectures—enabling our customers to accelerate AIOps solutions in networks.
Visit the Dell Telecom Multicloud Foundation site to learn more about our solution.