Ultra Ethernet: Libfabric Resource Initialization

Introduction

Ultra Ethernet uses the libfabric communication framework to let endpoints interact with AI frameworks and, ultimately, with each other across GPUs. Libfabric provides a high-performance, low-latency API that hides the details of the underlying transport, so AI frameworks do not need to manage the low-level details of endpoints, buffers, or the underlying address tables that map communication paths. This makes applications more portable across different fabrics while still providing access to advanced features such as zero-copy transfers and RDMA, which are essential for large-scale AI workloads.

During system initialization, libfabric coordinates with the appropriate provider—such as the UET provider—to query the network hardware and organize communication around three main objects: the Fabric, the Domain, and the Endpoint. Each object manages specific sub-objects and resources. For example, a Domain handles memory registration and hardware resources, while an Endpoint is associated with completion queues, transmit/receive buffers, and transport metadata. Ultra Ethernet maps these objects directly to the network hardware, ensuring that when GPUs begin exchanging training data, the communication paths are already aligned for low-latency, high-bandwidth transfers.

Once initialization is complete, AI frameworks issue standard libfabric calls to send and receive data. Ultra Ethernet ensures that this data flows efficiently across Continue reading

Technology Short Take 188

Welcome to Technology Short Take #188! I’m back once again with a small collection of articles and links related to a variety of data center-related technologies. I hope you find something useful!

Networking

Security

You don’t need quantum hardware for post-quantum security

Organizations have finite resources available to combat threats, both by the adversaries of today and those in the not-so-distant future that are armed with quantum computers. In this post, we provide guidance on what to prioritize to best prepare for the future, when quantum computers become powerful enough to break the conventional cryptography that underpins the security of modern computing systems.  We describe how post-quantum cryptography (PQC) can be deployed on your existing hardware to protect from threats posed by quantum computing, and explain why quantum key distribution (QKD) and quantum random number generation (QRNG) are neither necessary nor sufficient for security in the quantum age.

Are you quantum ready?

“Quantum” is becoming one of the most heavily used buzzwords in the tech industry. What does it actually mean, and why should you care?

At its core, “quantum” refers to technologies that harness principles of quantum mechanics to perform tasks that are not feasible with classical computers. Quantum computers have exciting potential to unlock advancements in materials science and medicine, but also pose a threat to computer security systems. The term Q-day refers to the day that adversaries possess quantum computers that are large and stable enough to Continue reading

Use Additional BGP Paths for IBGP Load Balancing

I wrote about the optimal BGP path selection with BGP additional paths in 2021, and I probably mentioned (in one of the 360 BGP-related blog posts) that you need it to implement IBGP load balancing in networks using BGP route reflectors. If you want to try that out, check out the IBGP Load Balancing with BGP Additional Paths lab exercise.

Click here to start the lab in your browser using GitHub Codespaces (or set up your own lab infrastructure). After starting the lab environment, change the directory to lb/4-ibgp-add-path and execute netlab up.

Connect and secure any private or public app by hostname, not IP — free for everyone in Cloudflare One

Connecting to an application should be as simple as knowing its name. Yet, many security models still force us to rely on brittle, ever-changing IP addresses. And we heard from many of you that managing those ever-changing IP lists was a constant struggle. 

Today, we’re taking a major step toward making that a relic of the past.

We're excited to announce that you can now route traffic to Cloudflare Tunnel based on a hostname or a domain. This allows you to use Cloudflare Tunnel to build simple zero-trust and egress policies for your private and public web applications without ever needing to know their underlying IP. This is one more step on our mission to strengthen platform-wide support for hostname- and domain-based policies in the Cloudflare One SASE platform, simplifying complexity and improving security for our customers and end users. 

Grant access to applications, not networks

In August 2020, the National Institute of Standards (NIST) published Special Publication 800-207, encouraging organizations to abandon the "castle-and-moat" model of security (where trust is established on the basis of network location) and move to a Zero Trust model (where we “verify anything and everything attempting to establish access").

Continue reading

Arista EOS Hates a Routing Instance with No Interfaces

I always ask engineers reporting a netlab bug to provide a minimal lab topology that would reproduce the error, sometimes resulting in “interesting” side effects. For example, I was trying to debug a BGP-related Arista EOS issue using a netlab topology similar to this one:

defaults.device: eos
module: [ bgp ]
nodes:
  a: { bgp.as: 65000 }
  b: { bgp.as: 65001 }

Imagine my astonishment when the two switches failed to configure BGP. Here’s the error message I got when running the netlab’s deploy device configurations Ansible playbook:

Read more …

The RUM Diaries: enabling Web Analytics by default

Measuring and improving performance on the Internet can be a daunting task because it spans multiple layers: from the user’s device and browser, to DNS lookups and the network routes, to edge configurations and origin server location. Each layer introduces its own variability such as last-mile bandwidth constraints, third-party scripts, or limited CPU resources, that are often invisible unless you have robust observability tooling in place. Even if you gather data from most of these Internet hops, performance engineers still need to correlate different metrics like front-end events, network processing times, and server-side logs in order to pinpoint where and why elusive “latency” occurs to understand how to fix it.

We want to solve this problem by providing a powerful, in-depth monitoring solution that helps you debug and optimize applications, so you can understand and trace performance issues across the Internet, end to end.

That’s why we’re excited to announce the start of a major upgrade to Cloudflare’s performance analytics suite: Web Analytics as part of our real user monitoring (RUM) tools will soon be combined with network-level insights to help you pinpoint performance issues anywhere on a packet’s journey — from a visitor’s browser, through Cloudflare’s network, to your Continue reading

TCG058: Creating the Internet Layer That Should Have Been With Avery Pennarun

In this deep dive episode, we explore the evolution of networking with Avery Pennarun, Co-Founder and CEO of Tailscale. Avery shares his extensive journey through VPN technologies, from writing his first mesh VPN protocol in 1997 called “Tunnel Vision” to building Tailscale, a zero-trust networking solution. We discuss how Tailscale reimagines the OSI stack by... Read more »

NAN100: A Retrospective On 100 Episodes of Network Automation Nerds

Network Automation Nerds has reached a special milestone: episode 100! Eric Chou looks back on 5 years of conversations with network automation pioneers, practitioners, and visionaries. Drew Conry-Murray from the Packet Pushers joins Eric, along with online guest Ioannis Theodoridis, to find out why Eric started the podcast, his goals for all these conversations, a... Read more »

SwiNOG 40: When a Routing Control Functions Is Too Fresh

During integration testing, I find unexpected quirks in network devices way too often. However, that’s infinitely better than experiencing them in production (even after thoroughly testing stuff) while discovering that your peers don’t care about routing security, RPKI, and similar useless stuff.

For example, what happens if you define a new Routing Control Function (RFC) on Arista EOS and apply it to BGP routing updates in the same configuration session? You’ll find out in the Sorry We Messed Up (video) presentation Stefan Funke had at SwiNOG 40 (note: the bug has been fixed in the meantime).

Calico Whisker vs. Traditional Observability: Why Context Matters in Kubernetes Networking

Are you tired of digging through cryptic logs to understand your Kubernetes network? In today’s fast-paced cloud environments, clear, real-time visibility isn’t a luxury, it’s a necessity. Traditional logging and metrics often fall short, leaving you without the context needed to troubleshoot effectively.

That’s precisely what Calico Whisker’s recent launch (with Calico v3.30) aims to solve. This tool provides clarity where logs alone fall short. In the sections below, you’ll get a practical overview of how it works and how it fits into modern Kubernetes networking and security workflows.

If you’re relying on logs for network observability, you’re not alone. While this approach can provide some insights, it’s often a manual, resource-intensive process that puts significant load on your distributed systems. It’s simply not a cloud-native solution for real-time insights.

So are we doomed? No. Calico Whisker transforms network observability from a chore into a superpower.

What is Calico Whisker?

Calico Whisker is a free, lightweight, Kubernetes-native observability user interface (UI) created by Tigera and introduced with Calico Open Source v3.30. It’s designed to give you a simple yet powerful window into your cluster’s network traffic, helping you understand network flows and evaluate policy behavior in real-time.

In Continue reading

Nvidia Takes The Commanding Lead In Datacenter Ethernet Switching

Well, that didn’t take long. In April 2020, Nvidia completed its $6.9 billion acquisition of Mellanox Technologies for its InfiniBand and Ethernet switching, and a little more than five years and a GenAI boom later Nvidia has been crowned the leading revenue generator for Ethernet switching in the datacenter by IDC. …

Nvidia Takes The Commanding Lead In Datacenter Ethernet Switching was written by Timothy Prickett Morgan at The Next Platform.

PP078: Using Free Tools for Detection Engineering

You can build effective, scalable detection pipelines using free and open-source tools like Zeek, Suricata, YARA, and Security Onion. Today on Packet Protector we welcome Matt Gracie, Senior Engineer at Security Onion Solutions — the team behind the open-source platform used for detection engineering, network security monitoring, and log management. Matt has over 15 years... Read more »

HS112: Standardizing NaaS Service Definitions

Network-as-a-Service (NaaS) promises enterprises the ability to set up and configure connectivity and network security with a couple of clicks. But for NaaS to truly transform enterprise networking, one thing has been missing: standards. Enter Mplify (formerly the Metro Ethernet Forum), a non-profit focused on standardizing NaaS service definitions. Mplify’s CTO, Pascal Menezes, joins Johna... Read more »
1 … 4 5 6 7 8 … 3,816