Lenovo Confluent

Abstract

Modern HPC and AI infrastructures continue to grow in scale and complexity. Administrators need a management platform that can provision, monitor, secure, and maintain large-scale server fleets while reducing operational overhead and accelerating time-to-production.

Lenovo Confluent is an open-source cluster provisioning and management platform developed by Lenovo for large-scale HPC and AI environments. It provides a unified framework for hardware discovery, bare-metal provisioning, operating system deployment, lifecycle management, and automation. For GPU-accelerated environments, Confluent applies the same repeatable operating model to GPU nodes and enables deployment workflows to include GPU drivers and the networking software stack required by the target solution.

Accelerate deployment

Move from discovered hardware to a configured cluster through repeatable, automated workflows.

Operate consistently

Manage heterogeneous node roles, architectures, and operating systems from a common interface.

Scale with confidence

Use Collective Mode for active/active load distribution, high availability, and large fleets.

Why Lenovo Confluent?

Automated cluster provisioning

Discover, configure, boot, and deploy servers at scale, reducing repetitive node-by-node tasks.

Scalability and availability

Extend a single management interface across multiple active Confluent servers for high availability with Collective Mode.

Hardware lifecycle management

Centralize power, firmware, BIOS/UEFI/BMC settings, health, telemetry, and support data.

Flexible OS deployment

Support disk-based and diskless images using PXE, HTTP(s) boot, or virtual media, with customization throughout the deployment process.

Open and automation-ready

Use the web UI, CLI, REST API, and Ansible-based automation workflows.

Secure by design

Use TLS-secured communication, TPM2, Secure Boot, and SSH certificate-based management patterns.

HPC and AI ecosystem integration

Confluent provides the provisioning and infrastructure-management foundation for GPU-accelerated HPC and AI clusters. GPU servers are managed as defined node roles, enabling administrators to standardize software, operating system deployments, and infrastructure configurations across the environment. Through operating system images, automated deployment workflows, configuration management, and integration with external tools such as Ansible, Confluent can be used to deploy and maintain the software components commonly found in modern HPC and AI platforms.

• GPU driver and software: Deploy NVIDIA CUDA and AMD ROCm software stacks, including data center GPU drivers and supporting packages, as part of disk-based or diskless operating system definitions. Groups and reusable node definitions help maintain a consistent GPU software environment and configuration across the cluster.
• Network driver: Deploy supported networking drivers and software, including NVIDIA DOCA-OFED for InfiniBand and RoCE environments, as well as the Cornelis Omni-Path software stack.
• Operating systems and hypervisors: Confluent supports the deployment of Red Hat Enterprise Linux, Rocky Linux, Ubuntu, AlmaLinux, Oracle Linux, SUSE Linux Enterprise Server, and Debian, as well as virtualization platforms such as VMware ESXi and Proxmox. Operating systems can be deployed in disk-based or diskless configurations across supported x86 and Arm platforms.
• High-Performance storage: Confluent can be used to deploy and configure environments based on leading parallel file systems, including IBM Storage Scale, Lustre, and BeeGFS, helping ensure consistent configuration of storage clients, servers, and supporting services.
• Observability and monitoring: Confluent can bootstrap commonly used monitoring and observability platforms, including Icinga2, Grafana, Prometheus, Alertmanager, and Loki. These tools complement Confluent's native hardware-health and telemetry capabilities with broader infrastructure, metrics, alerting, and log visibility.
• Scheduling and resource orchestration: Confluent can prepare the infrastructure and system foundation for environments using NVIDIA Slurm, NVIDIA Slinky, Siemens PBS Professional, Siemens Grid Engine, IBM Spectrum LSF, and Kubernetes distributions.
• Energy monitoring and optimization: Confluent can support the deployment of EAR, the Energy Aware Runtime, alongside the selected scheduler or orchestration platform. EAR provides capabilities for monitoring energy consumption and optimizing cluster power efficiency.

Example HPC and AI platforms

Confluent provides the infrastructure foundation upon which higher-level HPC and AI platforms can be deployed. Examples include:

• A Slurm-based HPC cluster for research, engineering, simulation, and technical computing.
• A Kubernetes-based platform for AI training, inference, and containerized services.
• A hybrid HPC/AI environment combining Slurm and Kubernetes, including architectures that use NVIDIA Slinky.

Positioning note: These are solution examples rather than an exhaustive list of supported platforms. Confluent handles the hardware and operating-system foundation; scripts, configuration workflows, and Ansible can be used to bootstrap the selected platform layer.

Business and operational outcomes

Faster time-to-production

Standardize discovery, provisioning, and configuration from the first node to the full cluster.

Lower operational friction

Bring provisioning, hardware control, telemetry, consoles, and automation into one management plane.

Greater consistency

Apply reusable definitions across compute, GPU, login, storage, and other node roles.

Built for growth

Move from a standalone server to a resilient Collective architecture without changing the operational model.

About Lenovo

Lenovo is a US$83 billion revenue global technology powerhouse, ranked #153 in the Fortune Global 500, and serving millions of customers every day in 180 markets. Guided by its vision of “Smarter Technology for All”, Lenovo is executing a Hybrid AI strategy that spans Personal AI – one personal AI, multiple devices; and Enterprise AI – helping customers turn data into insights and value. This strategy is delivered through the Group’s commitment to world-class innovation and a full-stack AI portfolio, including devices, infrastructure solutions, software, and services. Lenovo is widely recognized for its operational excellence, with a global footprint spanning more than 20 R&D locations and a supply chain that includes more than 30 manufacturing sites across 10 markets.