Itsaso Igeregi

🚀 Iberia Takes Flight with AWS Countdown Premium  ✈️ Spain's flagship airline just migrated their business-critical website to AWS with virtually zero downtime using AWS Countdown Premium—and the results speak for themselves: 10x higher deployment frequency 40% shorter development cycles 38% lower error rates The stakes? Their primary sales channel serving millions of customers. The challenge? Coordinating 300+ developers across multiple workstreams with no room for error. When an unexpected auto-scaling issue hit during the overnight cutover, our 24/7 AWS Countdown Premium team connected them with an Amazon EKS expert in minutes—keeping the migration on track. Now Iberia is accelerating AI adoption with Amazon Bedrock, building a generative AI travel assistant for personalized customer experiences.  Read the full story: https://lnkd.in/geUX9ZrG #AWS #CustomerSuccess #CloudMigration #AWSCountdownPremium

27

Telefónica is scaling its cloud-native modernization across Spain and Brazil using Red Hat OpenShift. Learn how this collaboration is helping them build a flexible, future-ready foundation for telco innovation.

4

Building the foundation for enterprise-scale Kubernetes.🌍 Announced today at hashtag#KubeCon Europe: Broadcom is accelerating our collaboration with the cloud-native ecosystem. As a top-five long-term contributor to Cloud Native Computing Foundation (CNCF) projects, we are committed to making Kubernetes more reliable and secure for every platform engineer. Today’s key announcements: 🛡️ Velero Donation: The Velero project has been accepted into the CNCF Sandbox to ensure its future as a vendor-neutral standard for backup and recovery. ⚙️ VKS 3.6 Launch: Introducing VMware vSphere Kubernetes Service 3.6, with bring-your-own CNI, RHEL 9 integration, and automated upgrade readiness checks. 🤝 Ecosystem Expansion: Strengthening platform flexibility through new validated integrations with F5, Kong and Tigera. From upstream contributions to enterprise-grade tools, we are making Kubernetes easier to manage at scale. 🔗 Get the details: https://dy.si/wgQknG

3

1 comentario

More than proud and truly humbled to contribute to this article, which synthesizes João Cavalheiro’s Master’s Degree Dissertation at Instituto Superior Técnico and highlights an important step forward in modern data center networking. The work explores a critical challenge in cloud-scale data centers: the mismatch between Kubernetes service-level load balancing and network-level load balancing in the data center fabric — a gap that can lead to suboptimal paths, higher latency, and inefficient resource utilization. By deploying intelligent agents on each leaf switches, implemented and experimentally validated using Nokia SR Linux and interacting directly with the Kubernetes API, this solution makes the network Kubernetes-aware. The agents dynamically translate pod and service placement into ECMP-aware forwarding decisions, enabling the fabric to adapt in near real time to workload changes. This work is also a clear manifest of how network agents can actively influence traffic flows through weighted load balancing, based on the number of active Kubernetes PODs per node. Leveraging gRPC, gNMI, YANG and the modern NDK (NetOps Development Kit), the solution demonstrates how programmability and automation can turn the data center fabric into an intelligent, workload-aware system. The results are very encouraging: • Negligible added delay compared to Kubernetes convergence times • Elimination of unnecessary inter-node redirection • Reduced latency and improved overall resource utilization Altogether, this work shows how tight integration between container orchestration platforms and data center network infrastructure can significantly improve efficiency, resilience, and performance in cloud deployments. I would like to sincerely thank everyone who contributed to this work. In particular, I’m grateful to my two colleagues at my time at Nokia Sérgio Santos and Helder Barata for their technical expertise, collaboration, and valuable insights throughout the project. A special thank you as well to the professor Rui Valadas at Instituto Superior Técnico, University of Lisbon, for their academic guidance, rigor, and continuous support, Last but not least, I would like to thank Tiago Amado for his leadership and valuable contributions in organising and guiding the quorum throughout the academic year. Congratulations to João Carvalheiro on an excellent and outstanding master’s dissertation, and thank you to all involved in making this collaboration a success. 🚀 🏄 Check this out: https://lnkd.in/dRNu5W4g #netops #agile #automation #srlinux #loadbalancer #containerlab #yang #grpc #gnmi #pygnmi #wecmp #bgp #k8s #cloud #kubeapi #metallb #cni #etcd

50

4 comentarios

Successful LLM training requires moving large amounts of data to the model's compute environment. In this excellent blog (https://lnkd.in/dzMx2rwd), Oracle networking expert Radulescu Marius cites #interconnection first among his key strategies for doing so successfully and shows how to use Equinix Fabric Cloud Router to move petabytes of data from AWS S3 to OCI Object Storage. It's a clear and comprehensive read on how to solve an increasingly common use case with an Equinix digital interconnection solution!

19

Don’t let OOMKilled surprise your team—treat memory management like mission-critical ops. This deep dive guide breaks down how OOMKills work—from Linux kernel decisions to QoS tiers—and maps to real-world fixes: setting memory requests/limits intelligently, tuning oomScoreAdj, scaling strategies, node-level planning, and app design tweaks (e.g. Java container support). For anyone running Kubernetes in production, these lessons help you avoid crashes, latency spikes, and costly interruptions. 😎 Awesome Author:SUBBA REDDY SANGHAM 📚 Read it here: https://lnkd.in/gKSMspQR #Kubernetes #DevOps #MemoryManagement #Reliability

11

Bare Metal Kubernetes: Is the performance advantage disappearing? For years, running Kubernetes on bare metal was the go-to for performance-obsessed teams. No virtualization layers. No shared resources. Just raw speed. But the landscape is changing. Today’s cloud-native tooling, container runtimes, and orchestration improvements are closing the gap. Virtualized environments are now offering performance that's nearly indistinguishable—while still giving you scalability, flexibility, and manageability. CTOs and engineering leaders: It may be time to rethink that bare-metal loyalty. Unless you're running ultra-low latency workloads (think: trading platforms or real-time AI inference), the performance trade-off may no longer be worth the operational complexity. 📉 The marginal gains are shrinking. 📈 The cost of complexity is growing. What’s the real ROI of squeezing out the last 5% of performance—if it costs you 50% more in engineering overhead? 👉 Full breakdown here: https://lnkd.in/eYPACz5Q #Kubernetes #CloudNative #DevOps #CTOInsights #InfrastructureEngineering #PlatformEngineering #CloudComputing #PerformanceEngineering #SRE #BareMetal #TechStrategy #DigitalTransformation

6

6 comentarios

When you scale with Karpenter, you gain speed and flexibility in your EKS cluster. The challenge: monitoring nodes that don’t rely on auto-scaling groups. 👉 At Teracloud, we built a Terraform module that automates the creation and deletion of CloudWatch alarms for every EC2 instance managed by Karpenter. Here’s what you get: ✅​ Proactive visibility into scale events ✅​ Less manual effort, more reliability ✅​ Seamless integration with your EKS environment With this open and adaptable solution, you turn complexity into operational simplicity. 📖 Read the blog and see how to optimize your Karpenter monitoring: https://lnkd.in/d-rFjPYw

7

For years, software systems were optimized for speed and efficiency. Low latency, fast response times, and high throughput were seen as the ultimate indicators of success. Today, the real differentiator is resilience. In distributed, cloud-native architectures, failure is not an edge case. Networks time out. Services crash. Dependencies slow down. Traffic spikes unexpectedly. The question is no longer if something will fail — it is how well your system handles it when it does. Modern architecture begins with an important assumption: failure is inevitable. Designing for failure means building systems that degrade gracefully instead of collapsing. It means implementing timeouts, retries with backoff strategies, circuit breakers, and fallback mechanisms. It means isolating services so that one failure does not cascade across the entire platform. Resilience is also about visibility. Observability — through logs, metrics, and distributed tracing — must be built into the architecture from the start. If you cannot detect issues quickly or understand their impact, recovery becomes guesswork. Another key element is proactive testing. Techniques such as chaos engineering intentionally introduce controlled failures to uncover weaknesses before they appear in production. This approach shifts resilience from reactive firefighting to deliberate design. AI can assist in generating infrastructure code or configuring services, but it cannot determine acceptable risk levels, business trade-offs, or how systems should behave under stress. Those decisions require context, judgment, and experience. As systems grow more distributed and interconnected, resilience becomes a core architectural principle rather than a secondary concern. Performance attracts users. Resilience keeps them. #SoftwareArchitecture #DistributedSystems #SystemDesign #ResilienceEngineering #SoftwareEngineering Ionidea post Likith S, Wed 12:46 Ok

27

⚠️ Running #ApigeeHybrid in a single region? Don't compromise on #HighAvailability! Learn essential #Cassandra best practices from Carlos Frias at Google Cloud. Discover how leveraging snitches and #Kubernetes integration can mitigate node failures and optimize performance. A must-read for #APIops and #SRE teams. #DataStax #Resilience | Dive in 👉🏼 https://google.smh.re/4tfN

1

Submarine cable cuts in the Red Sea disrupt connectivity between Europe and Asia. Latency spikes as traffic is re-routed around the broken cables. Kentik's Cloud Latency Map captures the impacts here: https://lnkd.in/eW6XMxae

252

2 comentarios

OpenStack can’t stop, won’t stop. Infrastructure investment has never been higher and all of that physical infrastructure needs software, and that’s what open infrastructure software is all about. OpenStack has long powered critical infrastructure and AI is just the latest reason.

72

3 comentarios

Our 4th publication in gNMI series for SONiC NOS: "SONiC Certificates Deployment Automation"! In this article we show how automation with Ansible simplifies the deployment of TLS certificates across SONiC devices - eliminating manual errors and simplifying management at-scale. Whether you're running a lab or tackling production environments, this guide provides practical playbook and insights to help you streamline certificate operations. Check it out and accelerate your SONiC gNMI journey! https://lnkd.in/eC8RrWrj #SONiC #gNMI #NetworkAutomation #Cisco

27

1 comentario

Given the scale and complexity of our infrastructure, at Google we have to strike a delicate balance between several competing priorities, including security, reliability, efficiency, velocity, and maintainability. Here are the three core pillars that define how we protect production workloads at Google-scale.

2

🔒 Impressive case study from Moeve (formerly Cepsa) on controlling cloud resource deployment at scale! This energy company with 90+ years of experience and 11,000+ employees is leveraging AWS CloudFormation Guard Hooks to: • Enforce governance policies across deployments • Control resource deployment at enterprise scale • Accelerate their digital transformation journey • Support Europe's energy transition goals Key insights: 🔹 How Guard Hooks provide automated policy enforcement 🔹 Real-world implementation in a large enterprise 🔹 Balancing innovation with governance requirements Great example of how cloud governance can enable rather than hinder innovation! 🌱 #AWS #CloudFormation #GuardHooks #CloudGovernance #EnergyTransition #DigitalTransformation #Moeve #Cepsa #EnterpriseCloud #DevOps #CloudSecurity #AWSBlog

1

Telefónica is scaling its cloud-native modernization across Spain and Brazil using Red Hat OpenShift! Learn how this collaboration is helping them build a flexible, future-ready foundation for telco innovation.

13

Announcing the General Availability of Holodeck 9.0 Revolutionize Your VCF Testing and Training! We're thrilled to announce the General Availability of Holodeck 9.0, your ultimate toolkit for deploying standardized and automated nested VMware Cloud Foundation (VCF) environments! Today more than ever, the ability to quickly and reliably test new technical capabilities is top of mind. While deploying nested VCF environments has always been possible, it often involves time-consuming manual configurations and significant hardware overhead. Holodeck changes the game by providing a self-contained, automated, and repeatable solution that dramatically reduces complexity and resource requirements. - Significantly Reduced Hardware Requirements: Holodeck virtualizes the entire VCF management domain and workload domains, allowing you to run a complete VCF environment on a single ESX host or a vSphere cluster with sufficient capacity. A typical VCF 9.0 deployment (four node management domain and three node VI workload domain) can be nested with approximately 30 CPU cores, 325 GB memory, and 1.1 TB disk. - Self-Contained Services: No more reliance on external data center services during your testing! Holodeck comes with built-in common infrastructure services like NTP, DNS, BGP, and DHCP, making each environment truly independent and requiring only a single external IP. - Isolated and Conflict-Free Networking: Eliminate the need for complex VLAN and BGP connections in your network during early testing phases. Holodeck can provide isolated networking, preventing conflicts with existing configurations and allowing for multiple, completely self-contained nested environments. - Automation and Repeatability: Deploying nested VCF environments with Holodeck is nearly hands-off and easily repeatable. This means your teams can spin up new testing environments quickly and consistently, accelerating your understanding and adoption of VCF's powerful features across various use cases. Holodeck is your ideal partner for gaining a deeper understanding of how VCF functions. Holodeck 9.0, supports deploying both VCF 5.2.x and VCF 9.0 environments. Download the Holodeck binaries and access the comprehensive documentation here: Binaries: https://lnkd.in/dRNndSiZ Documentation: https://lnkd.in/dPAcQNxN Note: The use of HoloDeck and the Offline depot appliance will require a download token which is produced with entitlement to VCF through one of the following ways: purchased, trial, NFR, pass the VCP, or pass the VCP + VMUG advantage #VCF9 #Holodeck #VMware #Broadcom #vExpert #homelab

30

Rundeck & Ansible are a great combination of every ITOps stack, specially when you need to scale to non-Ansible experts. Here's why:

4