OpenInfra Summit 2020
| title | author | length |
|---|---|---|
| Run your Kubernetes cluster on OpenStack in Production | Anita Tragler, Franck Baudin, Ramon Acedo Rodriguez | 26m |
Shows how Kubernetes on OpenStack uses Kuryr/CNI, OVN, Neutron and Octavia to get bare‑metal datapath: SR-IOV and OVS‑DPDK are exposed to pods, provider… | ||
| Getting Started with Zuul | James E. Blair | 25m |
Shows how to bootstrap a real Zuul instance from the same container images used in production by running docker-compose to start Zookeeper, nodepool… | ||
| E2E 5G network slicing in ONAP | Seshu Kumar Swaminathan, Lin Meng | 21m |
ONAP proves an end-to-end slicing pipeline: the Frankfurt release implements a core NSMF that accepts slice requests and creates a Network Slice Instance (NSI)… | ||
| Deploying Container Network Functions (CNF) on Kubernetes | AJAY KALAMBUR, Yichen Wang | 24m |
Deploys a Vector Packet Processing (VPP) CNF on Kubernetes 1.18 using Multus, SR‑IOV CNI and the SR‑IOV device plugin with DPDK passthrough; cluster is Ubuntu… | ||
| A Universe from Nothing: Containerised OpenStack deployment using Kolla, Ansible and Kayobe | Stig Telfer, John Garbutt, Mark Goddard | 40m |
Hands-on demo of building a production-like OpenStack control plane using Kolla, Kolla-Ansible and Kayobe: images pull from a local registry, containers are… | ||
| OpenStack on rails: meet MicroStack | Tytus Kurek | 20m |
MicroStack claims to be the first OpenStack “on rails” packaged as a snap-based single archive that installs across 40+ Linux distributions and boots with two… | ||
| Docker for Machine Learning! | Haseeb Niazi, Paniraja Koppa | 24m |
Builds and containerizes a complete ML pipeline: trains SVM and Random Forest on the Iris dataset, pickles the models, exposes prediction endpoints with Flask… | ||
| Magma 101: An introduction to Magma | Phil Ritter, Ulas Kozat | 18m |
Magma is a distributed, cloud‑native wireless data core that delivers a converged 4G EPC and managed Wi‑Fi today with 5G services under development; its three… | ||
| Cloud-hypervisor: A new choice for virtual machine monitor | Michael Zhao, Henry Wang | 24m |
Cloud-hypervisor is a Rust-based VMM built on rust-vmm crates and targets cloud workloads on x86_64 and ARM64, dropping legacy devices in favor of a small set… | ||
| So, you want to run Windows workloads on OpenStack? | Miklos Cari | 20m |
Production cloud shifts to 70% Windows VMs after COVID and runs Windows on KVM/QEMU with nested Hyper‑V working inside a Windows Server 2016 guest. Image build… | ||
| OpenStack Cluster Installer: the Debian way to deploy OpenStack | Thomas Goirand | 19m |
A Debian-packaged installer that automates bare-metal to running OpenStack, provisioning a PXE-booted Debian live discovery image, collecting… | ||
| Live migration from VMware to OpenStack | Jack Lee, Brin Zhang, Ethan Gao | 23m |
Claims a non‑intrusive live migration from VMware to OpenStack that preserves running services by combining agentless disk read via Watson Disk Developer Kit… | ||
| MicroStack: Putting OpenStack at the edge | Tytus Kurek | 17m |
OpenStack at the edge packaged as a snap-based OpenStack called MicroStack that installs with a single-command install and integrates with MAAS (bare metal)… | ||
| Ceilometer, CloudKitty and Gnocchi: a dynamic and agnostic cloud monitoring and billing stack. | Rafael Weingartner | 19m |
Deploys Ceilometer + Gnocchi + CloudKitty to capture, store and rate heterogeneous cloud usage, adding a dynamic pollster via YAML so new JSON APIs can be… | ||
| Mirantis OpenStack on Kubernetes | Artem Andreev | 24m |
Mirantis OpenStack on Kubernetes (MOSK) packages OpenStack services as containers, managed by Kubernetes operators and delivered on Docker Enterprise Container… | ||
| Achieve a single Openstack deployment in multiple datacenter | Victor Coutellier | 18m |
Société Générale stretches a single OpenStack region across two Paris datacenters (no L2, <5 ms latency) by deploying Kolla-Ansible with per-AZ group_vars… | ||
| Hassle-free migration from OVS to OVN | Frode Nordahl | 21m |
Procedure for switching Neutron from ML2+Open vSwitch to OVN with concrete steps and timings. Requires upgrading Neutron (Ussuri), enabling the Open vSwitch… | ||
| Own your YAML: extending Kustomize via Plugins | Matt McEuen | 17m |
Kustomize runs replacement transformer, templater, and sops plugin extensions to modify or generate Kubernetes YAML during render. Replacement copies specific… | ||
| Lessons Learned from a Large Scale OpenStack Deployment with Tripleo | Pradipta Sahoo, Sai Sindhur Malleni | 27m |
TripleO deploys an OpenStack Ussuri overcloud of 700+ compute nodes using one bare-metal undercloud, Ironic/PXE provisioning, three monolithic controllers and… | ||
| I Deployed my Snowflake and it Didn't Melt! Now What? | Gianpietro Lavado, Mark Beierl | 21m |
Deploys a full Magma EPC stack in minutes using OSM as orchestrator, with the EPC manager as a CNF on Kubernetes and access gateways as VMs on OpenStack, an… | ||
| Project Overview - StarlingX 101 | Yong Hu | 13m |
StarlingX presents a production-ready edge cloud stack that claims hardened Kubernetes certified by CNCF and optional containerized OpenStack deployments. It… | ||
| Time-Sensitive Networking (TSN) Enabling on StarlingX | Yi Wang | 24m |
StarlingX runs a customized kernel and container image with the TSN stack and passes an Intel I210 NIC into Kata containers to enable 802.1Qbv time-aware… | ||
| Cloud Hypervisor and Kata Containers: A path towards modern virtualization | Jose Carlos Venegas Munoz | 13m |
Kata Containers is integrated with Cloud Hypervisor to run each container inside its own VM using the rust‑vmm stack and aims for minimal emulation, low… | ||
| OpenStack COA Info Session | Paul Quigley | 26m |
Official COA is a hands‑on, vendor‑neutral practical exam that gives a blank Ubuntu 18.04 lab per candidate and asks to complete 20+ tasks (create… | ||
| Airship 101 | Drew Walters, Andrew Karandjeff | 13m |
Airship 2.0 presents a concrete redesign: it runs production 5G workloads at AT&T already with an alpha in May 2020 and beta in Sep 2020 and targets GA… | ||
| Cisco 5G Automated Deployment | Abdullah Saeed, Jeroen Wittock | 24m |
Demo shows Cisco automates a full 5G standalone packet core deployment using CVM with OpenStack to provision a seven‑node project, build VMs and a Kubernetes… | ||
| 10 years of OpenStack at CERN (From 0 to 300k cores) | Belmiro Moreira | 19m |
CERN describes building a production OpenStack cloud since 2013 that scales to hundreds of thousands of cores, with concrete wins and pain points: deployment… | ||
| Huawei’s Practice in Migration to Cloud Native based 5G Telco Cloud | Tony Ruan | 24m |
Thank you to Huawei for sponsoring the Open Infrastructure Summit!
This session is to discuss the main trends and practice of Telco Cloud migration in 5G… | ||
| Auto-scaling in Openstack without Telemetry | Dat Vu Tuan, Kien Nguyen-Tuan | 16m |
Explains why OpenStack Heat plus Telemetry fails in production: Gnocchi/Aodh development stalled since Rocky, Ceilometer only collects OpenStack metrics… | ||
| The amphorae of application delivery & security with Octavia | Saurabh Sureka, Richu Channakeshava, Hunter Thompson | 26m |
Production-ready Octavia amphorae run as VM, bare metal or container with an A10 provider driver that automates layer‑4 and layer‑7 load balancing, SSL offload… | ||
| Cisco 5G Automated Deployment | Jeroen Wittock, Abdullah Saeed | 24m |
Cisco demonstrates a full automated 5G standalone packet core deployment using Cisco Virtualized Infrastructure Manager and UltraCloud: CVM creates OpenStack… | ||
| Zuul at Volvo Cars | Johannes Foufas, Albin Vass | 15m |
Volvo Cars runs ZUUL CI across about 111 developers and claims a continuous green master since April 2018 by using Zuul v3 features like speculative merge and… | ||
| How to split a cake? GPUs with OpenStack | Sylvain Bauza | 24m |
Shows a working OpenStack setup that slices heterogeneous GPUs into multiple instance types: an Intel iGPU and an NVIDIA T4 are configured with kernel modules… | ||
| Connecting ecosystems: How Cinder CSI, Ember CSI & Manila CSI leverage OpenStack bits in Kubernetes | Christian Schwede, Gorka Eguileor, Eric Harney | 25m |
Cinder CSI uses Nova and Cinder to provision and attach volumes and supports topology, snapshots, cloning, inline raw block, expansion and volume cloning; it… | ||
| Combining Ansible and Terraform for CI – better together love story based on OVN-CI project | Szymon Datko, Arie Bregman | 20m |
A CI workflow that uses Terraform to provision OpenStack test nodes and produce a dynamic Ansible inventory (example: two instances) and Ansible to configure… | ||
| Updating Firmware Using Ironic And Redfish In The Datacenter And At The Edge | David Paterson, Chris Dearborn | 20m |
Ironic uses Redfish simple update to push firmware images over HTTP/HTTPS to BMCs and is validated on Dell PowerEdge R640, R6515 and R630xd, updating iDRAC… | ||
| K8s controller and CRD patterns in Python - kuryr-kubernetes case study | Michał Dulko, Maysa de Macedo Souza, Sarka Scavnicka | 20m |
Kuryr-Kubernetes implements Neutron networking and Octavia load balancers for pods, storing state in Kubernetes CRs instead of a DB and avoiding double… | ||
| Managed Kubernetes Service for Public Cloud with Openstack and Kubernetes | Sa Pham | 17m |
Deployed a managed Kubernetes service for a Vietnamese public cloud that replaces OpenStack Magnum (cluster creation often 10–15 minutes and unstable) and… | ||
| Introduction of FaaS (Function as a Service) in starlingX | Sharath Kumar K, Poornima Nagaraju | 18m |
Integration of FaaS into starlingX builds a working prototype that runs short-lived functions on edge nodes using the platform’s service and lifecycle… | ||
| You know nothing, Jon Snow! OpenStack troubleshooting from a beginner's perspective. | Elena Lindqvist | 21m |
Three real incidents and exact fixes: a VM unreachable on port 1521 turns out to be the service stopped (verify with telnet or /dev/tcp and start the service)… | ||
| OpenStack Client 101: Becoming a Comandline Poweruser | Nils Magnus | 25m |
Compares direct API curl calls, the Python SDK and per-service CLIs, then teaches the unified OpenStack Client (OSC) workflow: how to install inside a Python… | ||
| Discover OpenStack's nerve with oslo.metrics: Have a robust private cloud on a large scale | Motomu Utsumi, Reedip Banerjee | 26m |
oslo.metrics extracts internal metrics from oslo libraries and sends them in Prometheus exposition format to Grafana so operators can see concrete RPC and DB… | ||
| Lessons learnt expanding Cambridge University's CSD3 Supercomputer with OpenStack | John Garbutt, Paul Browne | 27m |
Cambridge expands CSD3 using Ironic bare-metal provisioning on a ~1000‑node pool and Dell PowerEdge C6420 servers with Mellanox ConnectX‑6 (HDR100/50G… | ||
| Auto-scaling Jenkins Cluster on OpenStack Cloud Platform | Cong Ha Minh, Chien Pham Tuong | 15m |
A Jenkins cluster runs on OpenStack using the OpenStack Cloud plugin to auto-provision and revoke build slaves, with two masters in active/passive mode and… | ||
| Scaling Bare Metal Provisioning with Nova and Ironic at CERN | Arne Wiebalck, Belmiro Moreira | 18m |
CERN runs Nova plus Ironic to provision physical servers and now manages 5,000 nodes -> 10,000; scaling hits are concrete not conceptual. iSCSI deploys cause… | ||
| China Tower's CDN edge computing solution based on Intel servers and StarlingX | Jianqing Jiang, Junyu He, Liang Zhang | 19m |
China Tower opens near one million machine rooms to OTT CDN vendors and pilots an edge CDN on Intel x86 servers using StarlingX to provide isolated… | ||
| The Sandwich Situation: Ansible Modules for the OpenStack SDK | Nils Magnus, Artem Goncharov | 30m |
Ansible modules are implemented by calling the OpenStack SDK so provisioning is literally a playbook invoking modules like openstack.cloud.server and… | ||
| The Progress for Cloud Computing on Arm architecture | ChangBo Guo, Zhenyu Zheng, Kevin Zhao | 29m |
Linaro runs a free Arm64 developer cloud using OpenStack and Magnum to offer VMs, Kubernetes and diskless boot via Ironic+Cinder; work focuses on landing an… | ||
| Living on the Edge with DCN Networking | Anita Tragler, Bernard Cafarelli | 29m |
Two reference architectures show what’s been built and what actually works: a telco edge using SR-IOV production-ready fast datapath, dual-stack provider… | ||
| The practice and landing of Kata containers in Ant Group and Alibaba Group | Fupan Li, Wei Yang | 17m |
Ant/Alibaba run Kata on thousands of servers hosting what they call billions of service endpoints and use Kata for resource and fault isolation, not just… | ||
| Digital Sovereignty - Why Open Infrastructure Matters | Kurt Garloff, marius feldman | 25m |
Claim: digital sovereignty needs open, modular commodity layers rather than vendor-locked centralized clouds. GAIA‑X proposes a federated catalog of service… | ||
| Building the European Weather Cloud with Open source software : An end to end journey. | Vasileios Baousis | 26m |
European Weather Cloud runs two on‑site OpenStack clusters (initially Rocky then upgraded towards Ussuri), fronted by Ceph Nautilus storage and a federation… | ||
| DPDK performance optimization for edge computing scenario | haizhong qin, Chenjie Xu | 25m |
Shows concrete OpenStack + DPDK tuning: CPU pinning, hugepages, PMD binding, NIC offloads and affinity changes; host and VM flavors set for pure isolation and… | ||
| Towards Enclave-as-a-Container with Inclavare containers and Occlum | yuntong jin, Hongliang Tian, tianjia zhang | 25m |
Inclavare Containers implements an OCI-compliant enclave runtime that wires a containerd shim to enclave runtimes, integrates with Kubernetes runtimeClass, and… | ||
| Data Protection in OpenStack: More than just the Data | Robert Rother | 21m |
Trilio argues that protecting OpenStack requires more than VM backups: legacy agent-plus-media-server setups are costly, hard to scale and can consume an… | ||
| CI-CT-CD, Devops loop for future Telco Cloud | Fu Qiao | 24m |
China Mobile describes a CI‑CT‑CD loop that automates vendor lab integration through to commercial NFV cloud, covering tens of thousands of servers across… | ||
| CI and CD as a Service: A multi-vendor driven system | Sharath Rao, Abhijeet Singh, Avinash Raghavendra | 22m |
A blueprint for a CI/CD-as-a-service that stitches multiple vendor deliveries into a single release pipeline, preserving vendor autonomy while removing manual… | ||
| Casting light on (Bare) Metal: Meet the Ironic Prometheus Exporter! | Iury Gregory Melo Ferreira | 25m |
Ironic Prometheus Exporter (introduced in OpenStack Train) converts BMC sensor pulls into Prometheus text files per node so Prometheus scrapes temperature… | ||
| Security-by-construction: How to weave authorization into modern app stacks using Open Policy Agent | Tim Hinrichs | 27m |
Claims that authorization can be enforced pervasively by running Open Policy Agent next to services, using network proxies for transparent enforcement and… | ||
| NTT and KDDI Challenges for Sustainable Infrastructure Transformation | Yasufumi Ogawa, Toshiaki Takahashi, Hiroshi Tsuji | 25m |
Telecom operators move a vendor‑specific VNFM to OpenStack Tacker and report concrete results: KDDI runs Tacker for fixed BNGs on 16M subscribers, Tacker adds… | ||
| Private cloud price-performance analysis | Tytus Kurek | 24m |
Claims that Canonical’s Charmed OpenStack plus Ubuntu and vendor racks (Gigabyte Dell Supermicro Lenovo HPE QCT) cuts TCO versus AWS/Azure/GCP. Uses a 750‑VM… | ||
| The Big Decision when Adopting Openstack as a Private Cloud | Hind Naser | 15m |
Advice: expect three hard constraints when deploying OpenStack on-premise—long-term scalability and hardware flexibility, the operational complexity of… | ||
| Tackling operational and capacity utilization challenges at a large scale enterprise deployment | Bogdan Katyński, Silvano Buback, Imtiaz Chowdhury | 25m |
Workday runs >400K cores and ~7K hypervisors across 25 clusters (growing toward 10K) and addresses capacity by simulating placements with a JavaScript tool… | ||
| Performance optimization of Neutron server and Agent for large-scale scenario | haizhong qin, Yongfeng Du | 28m |
500-node Rocky-based scale test reveals DHCP and OVS-agent bottlenecks; DHCP agent accumulates message backlogs causing nova port activation timeouts, and OVS… | ||
| China Mobile Software and Hardware Integrated Portable Cloud Platform Test System | Zhang Xiao Guang, Zhiqiang Yu | 22m |
China Mobile builds a wheeled, pre‑installed hardware+software test kit that boots and configures racks remotely and already validates tens of thousands of… | ||
| Observability in Kata containers 2.0 | Bin Liu | 29m |
Kata Containers 2 ships built-in observability: Prometheus-format metrics from runtime, VMM and guest agent, plus a new kata-agent-ctl for in-VM debugging. The… | ||
| Yet Another Monitoring Solution? Why APImon is Different | Artem Goncharov, Nils Magnus | 32m |
APImon is an API‑centric monitor that runs real user scenarios as Ansible playbooks against an OpenStack cloud serving more than 100,000 cores and many PBytes… | ||
| The Magic of the Libvirt Driver – What to Do When You Invariably Hit "No Valid Host" | Stephen Finucane | 31m |
Libvirt (QEMU/KVM) is feature‑rich but causes the infamous scheduler “no valid host” when placement, filters and the driver disagree—examples: pinned CPUs used… | ||
| SLO's for Openstack - the API for Engineering Teams | Kit Merker, Joseph Sandoval | 23m |
OpenStack and Kubernetes run across five data centers with over 200,000 cores and the team shows a live demo that injects load with go-gena and measures CPU… | ||
| OpenStack Block Storage differs from traditional block storage | Tushar Gite | 17m |
Compares classic SAN appliances to a modular OpenStack Cinder implementation that exposes Cinder API, scheduler, volume and backup services and presents… | ||
| Kubernetes as a container platform for the Edge | Karim MANAOUIL, Adrien Lebre | 23m |
Experimental results on Grid5000 evaluate a vanilla Kubernetes with 1 master + 99 workers under increasing WAN RTT from 1–400 ms using cluster-loader to create… | ||
| Physical tenant separation: evaluation of concepts and proposed implementations | Laura Geisler | 15m |
Evaluates concrete options for making tenant workloads physically exclusive: full extra clouds (high overhead), OpenStack Ironic bare‑metal nodes (user manages… | ||
| User Report of a Zuul Installation in Production | Artem Goncharov, Nils Magnus | 24m |
Production Zuul runs on bare metal with containers on Fedora CoreOS and Podman, but Zookeeper and Nodepool move to CentOS after reboots kill clusters; the… | ||
| Kata * TEE = A Lego-like Two-way Sandbox for Seamless Security and Privacy | Kailun Qin | 20m |
Kata Containers combines lightweight VMs and TEEs to create a two-way sandbox that protects container secrets from both hostile tenants and compromised hosts… | ||
| Learnings from having Zero touch device Onboard service on StarlingX | Poornima Nagaraju, Sharath Kumar K | 24m |
Deploys Intel SDO as Kubernetes workloads on StarlingX and shows an end‑to‑end Secure Device Onboard flow that succeeds on the edge cloud. Covers concrete… | ||
| Enhancement and Optimization of computing, storage & network for bare metal with hardware offloading | Yajun Yang, Yunxiang Tao, yin rong | 19m |
China Mobile builds an “elastic bare metal” eCloud that moves OpenStack/QEMU/OVS datapath into hardware: a Roka compute server plus a hypercard with SoC+FPGA… | ||
| SCS: Large federated infrastructure for sovereignty | Kurt Garloff, Christian Berendt | 20m |
European GAIA-X effort delivers the Sovereign Cloud Stack as a modular, four opens open source implementation and a certifiable standard that binds together… | ||
| Project Overview: OpenInfra Labs 101 | Michael Daitzman | 07m |
OpenInfra Labs consolidates the Mass Open Cloud, NERC and Open Cloud Testbed to publish reproducible stacks and ship cloud in a box releases that capture… | ||
| Improvements to Identity Federation in Keystone | Kristi Nikolla | 15m |
During the Ussuri cycle Keystone adds two concrete federation features: expiring group memberships persisted in the DB with a configurable per‑IdP TTL that… | ||
| Hardware Automation Deployment and Test with Big Scale in ChinaMobile NFV cloud | Zhiqiang Yu, Zhang Xiaoguang | 21m |
Automated hardware integration system for China Mobile NFV cloud configures and validates 50,000 servers with plug-and-play capability using an auto platform… | ||
| Multi-Arch Support in OpenStack | Jeremy Freudberg, Rico Lin | 23m |
OpenStack now runs an upstream CI pipeline for arm64 called check-arm64 pipeline using roughly 40 Linaro-donated servers, running unit tests via tox/Python 3.8… | ||
| Hybrid Cloud Ecology of Tencent, Reconstructing the future digital infrastructure | Ruan He | 15m |
Tencent claims a production hybrid cloud in use since 2012 that runs most internal apps and glues private T-Stack and public clouds via a dedicated line… | ||
| Dependency Analytics | Shubham Mathur, Darshan Vandra | 20m |
A VS Code/Eclipse/JetBrains extension hosted on OpenShift that scans manifests and produces a stacked report: security issues (drawn from NVD), license… | ||
| Making a better world by edge computing infrastructure: the experience and thinking from FiberHome | Wang Hao, Linxiang Chen | 16m |
FiberHome builds an edge stack based on StarlingX + OpenStack + Kubernetes and ships OS Light 2.0.4, deployable as full sub-clouds or lightweight nodes. A… | ||
| Toward Next Generation Container Image | Yan Song | 16m |
Special thanks to our 2020 Open Infrastructure Summit sponsors for making the event possible: Canonical (Ubuntu), Huawei, VEXXHOST, Cisco, Tencent Cloud… | ||
| Containers within the CNTT reference model | Georg Kunz, Gergely Csatari | 29m |
CNTT publishes a technology‑agnostic Reference Model and a Kubernetes Reference Architecture (RA2) with concrete requirements — e.g. Kubernetes must run one of… | ||
| Building high efficient storage infrastructure for secure container on top of SPDK | Xiaodong Liu, Changpeng Liu | 20m |
Claim: SPDK vhost-user lets Kata Containers get direct block devices with over 2x IOPS using one CPU core versus host kernel stacks. Implementation uses SPDK… | ||
| Building containers is fun. Let's see how? | Arun Chaudhary | 24m |
Live demo shows Docker plus libcap and systemtap producing a reproducible hard rule: dropping all capabilities breaks httpd because it lacks CAP_SETUID… | ||
| How open source contribution has increased work efficiency of a French bank ? | Victor Coutellier | 19m |
Private cloud team moves from VMware to OpenStack, implements changes upstream instead of forking, and measures concrete gains: added tags support to Heat… | ||
| Practice and Thinking of Telecom-cloud Management Platform based on Cloud-native | Yan Zhang, Mingxuan Li | 27m |
China Unicom builds a Unified Cloud Management Platform (UCMP) that claims national‑level control of multi‑cloud, multi‑DC resources, managing physical… | ||
| Enhancement and optimization of new heterogeneous accelerators based on Cyborg | Brin Zhang, Shaohe Feng, wenping song | 23m |
Cyborg adds vGPU and NVMe SSD support, extends FPGA tooling (DPDK/programming and online program APIs) and implements online bind/unbind for accelerator cards… | ||
| Standalone yet at scale: Ironic take on High Performance Computing | Timothy Randles, Jacob Anders, Julia Kreger | 15m |
Los Alamos prototypes a standalone Ironic deployment that provisions bare metal at scale, aiming to turn weeks into minutes, manage mixed fleets from 32 to… | ||
| How does the Telco industry benefit from the CNTT/OPNFV outcomes? | Walter Kozlowski, Tomas Fredberg, Karine Sevilla, Sukhdev Kapur | 22m |
CNTT publishes a concrete reference model with mapped infrastructure and workload profiles plus a conformance test suite that operators are already requiring… | ||
| BYO Infrastructure: Who needs to run a control plane, anyway? | Bruce Basil Mathews | 23m |
Argues that the control plane can run in public clouds while on‑prem keeps compute and storage, using Kubernetes everywhere and manage updates in a public… | ||
| Effective migration from Newton to Queens with minimal downtime | Attila Szlovencsák, Viktor Schlaffer | 22m |
A 6‑month migration moves ~2,000 VMs and 500 TB from a vendor OpenStack to an internal Queens Ceph-based cloud using a custom Python toolset that orchestrates… | ||
| I Don’t Think This Means What You Think It Means: Red Herrings in OpenStack | Florian Haas | 21m |
Three concrete OpenStack gotchas and their fixes: Neutron HA router VRIDs are 8‑bit so VRID is 8‑bit (max 255 HA routers) — hit that hard limit and must spread… | ||
| DIMINET: On Demand Multi-site Connectivity for Edge Infrastructures | David Espinel | 23m |
DIMINET is a Python service that glues independent OpenStack Neutron instances using the neutron-to-neutron interconnection plugin and BGP-VPNs (GoBGP as the… | ||
| It's a Multi-Mesh World | Lee Calcote | 28m |
Service meshes proliferate and three open specs aim to tame them: SMI for a lowest-common-denominator API, SMPS for repeatable performance benchmarks, and… | ||
| Declarative chain to Kubernetes multi clusters for automation of HA workloads. | Alex Barchiesi, Matteo Di Fazio, Marco Lorini | 18m |
GARR builds an automated pipeline from bare metal to federated Kubernetes using MAAS, juju bundles, OpenStack and Kubernetes so a declarative Juju config spins… | ||
| Tackling security challenges with Airship | Alexander Hughes | 26m |
Airship demonstrates concrete fixes: container images live on quay.io and clair scan finds 199 CVEs in an Ubuntu 16.04 pegleg image versus 44 in 18.04, proving… | ||
| Day-3-Operations: New Region and New Services | Clemens Hardewig, Sebastian Wenner, Nils Magnus | 23m |
Public cloud expansion to a new Amsterdam region uses OpenStack with containerized control planes and TripleO, deploying ~2,000 devices (1,400 servers, 500… | ||
| The Coral Reef Cloud: Autoscaling, Reservation, and Preemption | John Garbutt, Steve Brasier, Pierre Riteau | 20m |
OpenStack Blazar reservations get a temporal lease so platforms can guarantee future capacity while the cloud fills gaps with preemptible instances that Blazar… | ||
| Back to the original mission: an open cloud operating system | Kurt Garloff, Christian Berendt | 21m |
Project runs with more than a dozen contributors and commercial partners and delivers an automated OpenStack-based SCS stack using KVM, Ceph, Open vSwitch/OVN… | ||
| Significance of Hardware Classification combined with Host Configuration Operator | Digambar Patil, John Williams, Sirisha Gopigiri | 14m |
Special thanks to our 2020 Open Infrastructure Summit sponsors for making the event possible: Canonical (Ubuntu), Huawei, VEXXHOST, Cisco, Tencent Cloud… | ||
| What Makes Open Source Communities Thrive Remotely? | Aeva Black | 23m |
Remote productivity drops because loss of social grooming breaks team cohesion; evidence: Dunbar-style limit around 150 people, face-to-face meetups about… | ||
| Elastic Secure Infrastructure (ESI): I Learned to Share my Hardware; You Can Too! | Tzu-Mainn Chen | 20m |
Builds a working hardware-leasing stack for the Mass Open Cloud: adds ironic node multi-tenancy by adding owner/lessee fields and policy rules, implements a… | ||
| OSF Edge Computing WG: Defining the Undefinable | Gergely Csatari, Ildikó Váncsa, David Paterson, Beth Cohen | 23m |
The working group publishes a second white paper that defines two reference architectures — centralized and distributed control planes — and maps them to… | ||
| “I don’t qualify” and other tales: impact of hiring processes on diversity on open source | Meghan Heisler, Maari Tamm, Amy Marrich | 21m |
OpenStack-run internships (Google Summer of Code and others) place 55 interns over time with at least 15 from diverse geographies, use Gerrit patch history and… | ||
| Making the Business Case for Contributing to Open Source | Tobie Langel | 23m |
Companies actually use open source heavily — a 2018 Black Duck audit finds 96% of apps contain open source and open source often makes up over half a codebase… | ||
| nothing matches | ||