0

Dual-Stack Common-Services VRF Confuses Aruba CX

As I was running the netlab pre-release integration tests, I noticed that ArubaCX failed the IPv6 Common Services test (it worked before). Here’s the gist of what that test does:

• It creates three VRFs (red, blue, and common)
• It imports routes from red and blue VRF into the common VRF and routes from the common VRF into the red and blue VRF (the schoolbook example of common services VRF)
• Just to be on the safe side, it imports red routes into the red VRF and so on.

Here’s the relevant part of the netlab lab topology:

0

Sizing Up AWS “Blackwell” GPU Systems Against Prior GPUs And Trainiums

This week, Amazon Web Services announced the availability of its first UltraServer pre-configured supercomputers based on Nvidia’s “Grace” CG100 CPUs and its “Blackwell” B200 GPUs in what is called a GB200 NVL72 shared GPU memory configuration. …

Sizing Up AWS “Blackwell” GPU Systems Against Prior GPUs And Trainiums was written by Timothy Prickett Morgan at The Next Platform.

0

Will Companies Build Or Buy Their GenAI Models?

One of the biggest questions that enterprises, governments, academic institutions, and HPC centers the world over are going to have to answer very soon – if they have not made the decision already – is if they are going to train their own AI models and the inference software stacks that make them useful or just buy them from third parties and get to work integrating AI with their applications a lot faster. …

Will Companies Build Or Buy Their GenAI Models? was written by Timothy Prickett Morgan at The Next Platform.

0

AI Metrics with InfluxDB Cloud

The InfluxDB AI Metrics dashboard shown above tracks performance metrics for AI/ML RoCEv2 network traffic, for example, large scale CUDA compute tasks using NVIDIA Collective Communication Library (NCCL) operations for inter-GPU communications: AllReduce, Broadcast, Reduce, AllGather, and ReduceScatter.

The metrics include:

• Total Traffic Total traffic entering fabric
• Operations Total RoCEv2 operations broken out by type
• Core Link Traffic Histogram of load on fabric links
• Edge Link Traffic Histogram of load on access ports
• RDMA Operations Total RDMA operations
• RDMA Bytes Average RDMA operation size
• Credits Average number of credits in RoCEv2 acknowledgements
• Period Detected period of compute / exchange activity on fabric (in this case just over 0.5 seconds)
• Congestion Total ECN / CNP congestion messages
• Errors Total ingress / egress errors
• Discards Total ingress / egress discards
• Drop Reasons Packet drop reasons

This article shows how to integrate with InfluxDB Cloud instead of running the services locally.

Note: InfluxDB Cloud has a free service tier that can be used to test this example.

Save the following compose.yml file on a system running Docker.

configs:
  config.telegraf:
    content: |
      [agent]
        interval = '15s'
        round_interval = true
        omit_hostname = true
      [[outputs.influxdb_v2]]
        urls = ['https://<INFLUXDB_CLOUD_INSTANCE>.cloud2.influxdata.com']
         Continue reading

0

Quicksilver v2: evolution of a globally distributed key-value store (Part 1)

Quicksilver is a key-value store developed internally by Cloudflare to enable fast global replication and low-latency access on a planet scale. It was initially designed to be a global distribution system for configurations, but over time it gained popularity and became the foundational storage system for many products in Cloudflare.

A previous post described how we moved Quicksilver to production and started replicating on all machines across our global network. That is what we called Quicksilver v1: each server has a full copy of the data and updates it through asynchronous replication. The design served us well for some time. However, as our business grew with an ever-expanding data center footprint and a growing dataset, it became more and more expensive to store everything everywhere.

We realized that storing the full dataset on every server is inefficient. Due to the uniform design, data accessed in one region or data center is replicated globally, even if it's never accessed elsewhere. This leads to wasted disk space. We decided to introduce a more efficient system with two new server roles: replica, which stores the full dataset and proxy, which acts as a persistent cache, evicting unused key-value pairs to free Continue reading

0

How BigQuery Combines Data And AI For Business Transformation

AI has the power to transform how organizations derive insights, make decisions, and unlock value, but all that depends on the quality of the data. …

How BigQuery Combines Data And AI For Business Transformation was written by Timothy Prickett Morgan at The Next Platform.

0

With Money And Rhea1 Tapeout, SiPearl Gets Real About HPC CPUs

The European Union cannot practically declare its independence from Nvidia GPUs any more than any other nation can at this point. …

With Money And Rhea1 Tapeout, SiPearl Gets Real About HPC CPUs was written by Timothy Prickett Morgan at The Next Platform.

0

AWS Transit Gateway VPN Attachment and Resource Access Manager (VII)

In the previous post, we covered the basics of Transit Gateway, what it is, what problem it solves, and we also looked at how to create one. We walked through attaching two VPCs to the TGW and establishing connectivity between them. We also covered the important concepts of TGW attachments, associations, and propagations.

AWS Transit Gateway Introduction

This complexity (when you have many VPCs) is why, in this post, we will look at AWS Transit Gateway (TGW). A Transit Gateway is an incredibly important networking resource in AWS that solves these scaling challenges.

PacketswitchSuresh Vina

In this post, we will build on that knowledge and look at

• Creating and attaching a Site-to-Site VPN to the Transit Gateway.
• We will also cover sharing a single Transit Gateway across different AWS accounts using Resource Access Manager.

As always, if you find this post helpful, press the ‘clap’ button. It means a lot to me and helps me know you enjoy this type of content. If I get enough claps for this series, I’ll make sure to write more on this specific topic.

Transit Gateway AWS Site-to-Site VPN Attachment

We have already seen how to create a Site-to-Site Continue reading

0

Worth Reading: The Secret Rules of the Terminal

Did you ever wonder why pressing an up-arrow in a (Linux) terminal window sometimes recalls the previous command but other times creates ^[[A?

Julia Evans did, and spent months exploring the quirks of the Linux terminal (and writing blog posts describing what she found), finally resulting in The Secret Rules of the Terminal (including the various shells, terminal emulators, escape codes, and TTY driver). A must-read if you’re a newbie who wants to understand why things happen the way they do.

0

Only The Biggest Neoclouds Will Survive

There are definitely easier businesses to be in than operating a neocloud. …

Only The Biggest Neoclouds Will Survive was written by Timothy Prickett Morgan at The Next Platform.

0

How TimescaleDB helped us scale analytics and reporting

At Cloudflare, PostgreSQL and ClickHouse are our standard databases for transactional and analytical workloads. If you’re part of a team building products with configuration in our Dashboard, chances are you're using PostgreSQL. It’s fast, versatile, reliable, and backed by over 30 years of development and real-world use. It has been a foundational part of our infrastructure since the beginning, and today we run hundreds of PostgreSQL instances across a wide range of configurations and replication setups.

ClickHouse is a more recent addition to our stack. We started using it around 2017, and it has enabled us to ingest tens of millions of rows per second while supporting millisecond-level query performance. ClickHouse is a remarkable technology, but like all systems, it involves trade-offs.

In this post, I’ll explain why we chose TimescaleDB — a Postgres extension — over ClickHouse to build the analytics and reporting capabilities in our Zero Trust product suite.

After a decade in software development, I’ve grown to appreciate systems that are simple and boring. Over time, I’ve found myself consistently advocating for architectures with the fewest moving parts possible. Whenever I see a system diagram with more than three boxes, I ask: Why Continue reading

0

Calico Whisker & Staged Network Policies: Secure Kubernetes Workloads Without Downtime

Rolling out network policies in a live Kubernetes cluster can feel like swapping wings mid-flight—one typo or overly broad rule and critical traffic is grounded. Calico’s Staged Network Policies remove the turbulence by letting you deploy policies in staged mode, so you can observe their impact before enforcing anything. Add Whisker, the open-source policy enforcement and testing tool (introduced as part of Calico Open Source 3.30) that captures every flow and tags it with a policy verdict, and you’ve got a safety harness that proves your change is sound long before you flip the switch. In this post, we’ll walk you through how you can leverage these capabilities to tighten security, validate intent, and ship changes confidently—without a single packet of downtime.

Deploying a Kubernetes Cluster

Calico for Policy is a CNI agnostic tool. Refer to the Calico Open Source docs for a list of supported CNIs. The git repository for this blog post can be found here.

For this post, let’s deploy a simple AKS cluster with Azure CNI.

## Configure 
az group create --name calicooss --location eastus2

## Create a 3 node AKS cluster with Azure CNI
az aks create \
  --resource-group calicooss \
  --name  Continue reading

0

Silicon One: Many Cisco Chips With One Architecture Chasing Many AI Workloads

With AI being the biggest change in IT infrastructure since the Dot Com boom, it was no surprise that at the annual Cisco Live event last month in San Diego, the focus was on AI – and particularly agentic AI – and how the networking giant differentiates itself from other infrastructure vendors when it comes to the emerging technology. …

Silicon One: Many Cisco Chips With One Architecture Chasing Many AI Workloads was written by Jeffrey Burt at The Next Platform.

0

Introducing simple and secure egress policies by hostname in Cloudflare’s SASE platform

Cloudflare’s SASE platform is on a mission to strengthen our platform-wide support for hostname- and domain-based policies. This mission is being driven by enthusiastic demands from our customers, and boosted along the way by several interesting engineering challenges. Today, we’re taking a deep dive into the first milestone of this mission, which we recently released in open beta: egress policies by hostname, domain, content category, and application. Let’s dive right in! 

Customers use our egress policies to control how their organization's Internet traffic connects to external services. An egress policy allows a customer to control the source IP address their traffic uses, as well as the geographic location that their traffic uses to egress onto the public Internet. Control of the source IP address is especially useful when accessing external services that apply policies to traffic based on source IPs, using IP Access Control Lists (ACLs). Some services use IP ACLs because they improve security, while others use them because they are explicitly required by regulation or compliance frameworks. 

(That said, it's important to clarify that we do not recommend relying on IP ACLs as the only security mechanism used to gate Continue reading

0

Ultra Ethernet

Introduction

Remote Direct Memory Access over Converged Ethernet (RoCE) is a transport model that extends InfiniBand semantics over Ethernet networks. It enables direct memory access between hosts by encapsulating InfiniBand transport headers—such as the InfiniBand Transport Header (IBTH) and the RDMA Extended Transport Header (RETH)—within Ethernet, IP, and UDP packets. In by book "Deep Learning for Network Engineers" Chapter 9, describes how RDMA NICs process application work requests, known as InfiniBand verbs, and how these are encoded into IBTH and RETH headers for delivery to remote targets using RoCEv2.

This post shifts focus to the Ultra Ethernet Transport (UET) model, developed by the Ultra Ethernet Consortium (UEC). UET defines an alternative RDMA transport architecture that operates over standard Ethernet networks, without relying on InfiniBand message formats or semantics. While both RoCEv2 and UET enable remote memory access between nodes, UET is not based on InfiniBand transport headers, and the term RoCE is not used in UET systems.

Instead, UET introduces a new Ultra Ethernet (UE) layer composed of several sublayers, including the Semantic Sublayer (SES) and the Packet Delivery Sublayer (PDS). These sublayers are responsible for encoding and transmitting RDMA operations—such as memory addresses, remote keys (RKEYs), operation codes, and Continue reading