AI Infrastructure Fleet Management Made Easy with COSMOS

AI infrastructure demands large server deployments in hyperscale data centers to meet requirements for compute power, efficiency, and total cost of ownership (TCO). However, managing such a large fleet of systems presents complex challenges of observability, data collection, and fault isolation.

COnnectivity System Management and Optimization Software – or COSMOS – is a key component of Astera Labs’ Intelligent Connectivity Platform. It enables management and optimization of resources for large fleets at cloud-scale via link, fleet, and RAS management capabilities.

COSMOS Monitors and Manages a System or Fleet

Why Use COSMOS?

Gain deep visibility into the thousands of data links in your fleet to maximize system availability

Predict when and where failures might occur – and isolate them before they happen

Understand real-time link level performance, temperature and per lane signal quality

COSMOS Highlights

Operates in on-chip microcontrollers and systems baseboard/system management controllers
Provides link management, fleet management and reliability/availability/serviceability (RAS) features across the entire product portfolio
Delivers enhanced diagnostics and telemetry features with in-band and out-of-band management
Monitors the state and health of a systems’ critical data links, such as link rate, link stability, error rate, receiver margins, and more
Offers enhanced security features enabling device updates and configurations in a secure manner
Available as a C-SDK, Python-SDK and Reference Applications (or Platform Utilities)

Request the White Paper

Cloud Infrastructure Fleet Management Made Easy with COSMOS

Download

Combating Noisy Neighbors with Scorpio P-Series Fabric Switches

AI server designs are being impacted by an issue that becomes increasingly worse as GPUs scale to meet demands of AI workloads. The issue: noisy neighbors! Scorpio P-Series Fabric Switches – the industry’s first PCIe 6 fabric switch – are architected for mixed traffic AI head-node traffic connectivity (GPU-to-CPU/NIC/SSD). Let’s take a closer look at the problem of noisy neighbors…

Advancing CXL with Interoperable Solutions

In AI and cloud infrastructure, seamless connectivity and scalable memory expansion solutions are key to unlocking performance at scale. As AI workloads grow more demanding, the need for robust, standards-based and interoperable solutions is greater than ever. The CXL Consortium Compliance Program was established to validate end-products against the CXL Specification, ensuring they meet the…

Seven Key Innovations Shaping AI Connectivity Showcased at DesignCon 2025

Astera Labs will be at DesignCon 2025, taking place January 28-30 at the Santa Clara Convention Center, to showcase our latest chip, board, and system design innovations for AI and cloud infrastructure.Join us at Booth #755 to see our Intelligent Connectivity Platform of PCIe®, Ethernet, and CXL® connectivity solutions in action and learn how we are unleashing the full potential of…

Astera Labs Optimizes Connectivity for NVIDIA Blackwell-Based MGX Platforms at Scale

Seamless Scorpio Smart Fabric Switch integration with NVIDIA MGX™ platform delivers PCIe® 6-ready modular designs for rapid deployment across a range of AI serversSANTA CLARA, CA, U.S. – March 18, 2025 – Astera Labs, Inc. (Nasdaq: ALAB), a global leader in semiconductor-based connectivity solutions for AI and cloud infrastructure, today announced its Scorpio Smart Fabric Switches…

Astera Labs Expands Cloud-Scale Interop Lab Leadership to Propel Next-Gen PCIe 6.x Ecosystem

Comprehensive Cloud-Scale Interop Lab testing of Scorpio Smart Fabric Switches advances PCIe 6.x ecosystem enablement, fast-tracking customer platform designs, development, and time-to-marketSANTA CLARA, CA, U.S. – March 13, 2025 – Astera Labs, Inc. (Nasdaq: ALAB), a global leader in semiconductor-based connectivity solutions for AI and cloud infrastructure, today announced the expansion…

Astera Labs Appoints Dr. Craig Barratt to Board of Directors

Dr. Barratt brings leadership experience in scaling high-growth public and private technology companiesSANTA CLARA, CA, U.S. – March 3, 2025 – Astera Labs, Inc. (Nasdaq: ALAB), a global leader in semiconductor-based connectivity solutions for AI and cloud infrastructure, today announced the appointment of Dr. Craig Barratt to its Board of Directors. Dr. Barratt is a seasoned technology…

Astera Labs to Participate in the Morgan Stanley Technology, Media & Telecom Conference

SANTA CLARA, CA, U.S. – February 25, 2025 – Astera Labs (Nasdaq: ALAB), a global leader in semiconductor-based connectivity solutions for AI and cloud infrastructure, today announced that it will participate in the Morgan Stanley Technology, Media & Telecom Conference on March 4, 2025. Astera Labs’ presentation is scheduled for 4:05 pm PT. A webcast of the session will be made available…

End-to-End PCIe 6 Interop with NVIDIA Blackwell

In partnership with major hosts, GPU, network and storage ecosystem partners, we’re showing end-to-end interoperability with our Scorpio P-Series Fabric Switch, Aries Smart DSP Retimer, NVIDIA Blackwell GPU, 5th Gen Intel® Xeon® CPU, NVIDIA ConnectX-7 NIC, and a Micron PCIe 6 SSD for ecosystem enablement.

Optimizing NVIDIA MGX Platforms with Scorpio

At NVIDIA GTC 2025, Astera Labs is showcasing its collaboration with NVIDIA and Wistron. This is the first of many solutions using our Scorpio Smart Fabric Switches to deliver leading performance through a modular design that scales across different configurations for PCIe 6-ready NVIDIA Blackwell-based MGX platforms.For this demo, utilizing the PCIe 6-ready NVIDIA Blackwell-based MGX platform…

How We Test – Scorpio

Scorpio Smart Fabric Switches undergo interoperability and stress testing with a wide range of PCIe 6.x exercisers, analyzers, GPUs, CPUs, NICs, SSDs, and Switches across multiple PCIe generations and topologies.This rigorous testing enables customers to design with confidence, minimize interoperation risk, and reduce overall development time and costs.