Linaro has enabled server class workloads for ARM servers by optimizing key open source software. They have contributed patches to projects like the Linux kernel, KVM, Xen, OpenJDK, Hadoop, and OpenStack. This has allowed OpenStack to run on ARMv8 hardware, with all applicable Tempest tests passing. Linaro is also working on optimizations for server workloads like the LAMP stack, HDFS, and HipHop JIT. Their efforts are helping to accelerate ARM's adoption in the server market.
Is OpenStack Neutron production ready for large scale deployments?Елена Ежова
OpenStack Neutron with ML2 OVS has always been a challenging component in terms of performance and scalability. However, in recent releases, several enhancements and bug-fixes have resulted in significant improvements in overall reliability, performance and scalability of Neutron. In this presentation, we will share the results of our testing (both control-plane and data-plane) at large scale and provide a detailed data-driven analysis that explores the true scale limits and bottlenecks of Neutron.
Red Hat demo of OpenStack and ODL at ODL summit 2016 RedHatTelco
Red Hat demonstrated OpenDaylight (ODL) as an SDN Controller for OpenStack. We showed the integration of the Boron release of OpenDaylight with the Mitaka release of OpenStack. The primary objective of the demo was to show how NetVirt can easily create and manage virtual networks that are flexible, secure and scalable.
Is OpenStack Neutron production ready for large scale deployments?Елена Ежова
OpenStack Neutron with ML2 OVS has always been a challenging component in terms of performance and scalability. However, in recent releases, several enhancements and bug-fixes have resulted in significant improvements in overall reliability, performance and scalability of Neutron. In this presentation, we will share the results of our testing (both control-plane and data-plane) at large scale and provide a detailed data-driven analysis that explores the true scale limits and bottlenecks of Neutron.
Red Hat demo of OpenStack and ODL at ODL summit 2016 RedHatTelco
Red Hat demonstrated OpenDaylight (ODL) as an SDN Controller for OpenStack. We showed the integration of the Boron release of OpenDaylight with the Mitaka release of OpenStack. The primary objective of the demo was to show how NetVirt can easily create and manage virtual networks that are flexible, secure and scalable.
OpenDaylight: an open source SDN for your OpenStack cloudAnees Shaikh
Presented at the 2013 OpenStack Summit in Hong Kong.
Authors: Stephan Baucke, Kyle Mestery, Anees Shaikh, Chris Wright
OpenDaylight is an exciting new community-led, open source project focused on accelerating adoption of software-defined networking (SDN) by providing a robust SDN platform on which the industry can build and innovate. An OpenDaylight controller provides flexible management of both physical and virtual networks. The open source nature of the project and its flexible network management capabilities make it an ideal SDN platform to integrate with Neutron.
In this session, OpenDaylight community members from Cisco, IBM, RedHat, and Ericsson will describe the OpenDaylight project goals and platform architecture, as well as the roadmap and progress to date. OpenDaylight brings together a number of virtual networking approaches, and we will discuss integration approaches with OpenStack Neutron that provide flexibility for OpenStack administrators and users. Details of our initial Neutron integration will also be demonstrated for attendees.
Attendees will leave this session with a greater understanding of what OpenDaylight is, and how it can integrate with OpenStack Neutron to provide a powerful SDN-based networking solution for OpenStack Clouds.
Presentation delivered at LinuxCon China 2017.
Open vSwitch (OVS) is a multilayer open source virtual switch. OVS is designed to enable massive network automation through programmatic extension, while still supporting standard management interfaces. OVN is a new network virtualization project that brings virtual networking to the Open vSwitch user community. OVN includes logical switches and routers, security groups, and L2/L3/L4 ACLs, implemented on top of a tunnel-based overlay network.
In this presentation, we will provide an overview of the current state of the projects and their future plans, such as:
- The current state of the Linux, DPDK, and Hyper-V ports
- A status update on a portable BPF-based datapath
- The latest stateful and OpenFlow features available in OVS
- Performance and debugging enhancement to OVN
- OVN features under development such as ACL logging and encrypted tunnels
Improving Network Application Performance using Load Aware LibeventdevMichelle Holley
Compared to load unaware packet distribution mechanisms often used in the run to completion model, an event scheduler improves core utilization and better handles dynamic traffic mixes by scheduling packets to cores according to their load. It simultaneously provides both atomicity and packet ordering. Hardware-based event schedulers can also provide low-latency inter-core communication. The libeventdev library from Data Plane Development Kit (DPDK) helps developers leverage the event scheduler model.
About the presenter: Sundar Vedantham, Intel, is a Senior Technical Manager working in the Data Center Group in Allentown, PA. His research interests include network traffic and congestion management, high-speed networking, and theoretical computer models, areas in which he holds patents and has published papers, book chapter & articles. He received his Ph.D. in Computer Science in 1997 from Louisiana State University. He enjoys writing articles in English and Tamil to help improve public understanding of technical details behind the fields he has worked on and to attract young students to get into STEM fields.
VMware ESXi - Intel and Qlogic NIC throughput difference v0.6David Pasek
We are observing different network throughputs on Intel X710 NICs and QLogic FastLinQ QL41xxx NIC. ESXi hardware supports NIC hardware offloading and queueing on 10Gb, 25Gb, 40Gb and 100Gb NIC adapters. Multiple hardware queues per NIC interface (vmnic) and multiple software threads on ESXi VMkernel is depicted and documented in this paper which may or may not be the root cause of the observed problem. The key objective of this document is to clearly document and collect NIC information on two specific Network Adapters and do a comparison to find the difference or at least root cause hypothesis for further troubleshooting.
Opensource approach to design and deployment of Microservices based VNFMichelle Holley
Microservice is gaining increased adoption in the Telco NFV world. It is key to understand the design and deployment methodologies involved in developing Microservice based VNF. This talk provides an opensource practitioner approach to building and deploying a Microservice based VNF and includes the following: - Design patterns, workflow models - Design models for VNF placement, capacity management, scale-in/out and resiliency - Deployment considerations that includes handing of scale and fault tolerant VNF using well known Opensource tools.
About the presenter: Prem Sankar works for Ericsson Opensource Ecosystem team and part of the Opendaylight and OPNFV team in Ericsson. Prem evangelizes SDN and Cloud and has given many sessions and conducted workshops around SDN and ODL. Prem is PTL of ODL COE project and currently driving the Kuberenetes and ODL Integration in Opendaylight community. Prem is a frequent speaker at opensource summits and has presented in Opendaylight, OPNFV and Open networking summits.
HKG15-110: ODP Project Update
---------------------------------------------------
Speaker: Bill Fischofer
Date: February 9, 2015
---------------------------------------------------
★ Session Summary ★
This session provides a summary of ODP activities since LCU ‰Û÷14 and highlights the main features of ODP v1.0 for applications as well as the validations used by conforming ODP implementation.
--------------------------------------------------
★ Resources ★
Pathable: https://hkg15.pathable.com/meetings/250771
Video: https://www.youtube.com/watch?v=xABcGPOCOuU
Etherpad: http://pad.linaro.org/p/hkg15-110
---------------------------------------------------
★ Event Details ★
Linaro Connect Hong Kong 2015 - #HKG15
February 9-13th, 2015
Regal Airport Hotel Hong Kong Airport
---------------------------------------------------
http://www.linaro.org
http://connect.linaro.org
Looking for a way to deploy a stable OpenStack Cloud Environment with Opendaylight at ease? This session is about learning to deploy a Cloud environment with OPNFV Fuel deployer. Fuel is a deployment tool which deploys a wide variety of distributions with third party plugins like OpenDayLight, while abstracting out complexities of the deployment. The intent of this session is to familiarize deployment of OpenStack with OpenDaylight.
About the author: Pramod is a software developer in OpenStack and OpenDayLight, working for OTC, SSG at Intel. His Area of Interest is in Cloud Networking and Applications. He has prior experience in Databases and his current focus is on developing features of Cloud Networking Platform. He holds Masters Degree from San Jose State University.
June Boston openStack Summit: Preparing quantum for the data centerKamesh Pemmaraju
Quantum, OpenStack's virtual networking service, is slated to replace Nova's networking service in the upcoming Folsom release. Initial Quantum development has focused on the basic requirements of a green-field Nova cloud deployment. Additional requirements will be discussed that enable Quantum to support private Nova cloud deployments within existing data centers, as well as enterprise virtualization with the oVirt project. Initial work is underway for Folsom, and additional work will follow.
LCU14 310- Cisco ODP
---------------------------------------------------
Speaker: Robbie King
Date: September 17, 2014
---------------------------------------------------
★ Session Summary ★
Cisco to present their experience using ODP to provide portable accelerated access to crypto functions on various SoCs.
---------------------------------------------------
★ Resources ★
Zerista: http://lcu14.zerista.com/event/member/137757
Google Event: https://plus.google.com/u/0/events/ckmld1hll5jjijq11frbqmptet8
Video: https://www.youtube.com/watch?v=eFlTmslVK-Y&list=UUIVqQKxCyQLJS6xvSmfndLA
Etherpad: http://pad.linaro.org/p/lcu14-310
---------------------------------------------------
★ Event Details ★
Linaro Connect USA - #LCU14
September 15-19th, 2014
Hyatt Regency San Francisco Airport
---------------------------------------------------
http://www.linaro.org
http://connect.linaro.org
OpenStack is a free and open-source software platform for cloud computing, mostly deployed as an infrastructure-as-a-service (IaaS). OpenDaylight is an open source project under the Linux Foundation with the goal of furthering the adoption and innovation of SDN through the creation of a common industry supported platform.
In this session, I will talk about how OpenStack and OpenDaylight can be combined together to solve real world business cases and networking needs. We will cover:
- What is OpenDaylight
- Use cases for OpenDaylight with OpenStack
- The OpenDaylight NetVirt project
- How OpenDaylight interacts with OpenStack
- The future of OpenDaylight, and how we see it help solving challenges in the networking industry such as NFV, container networking and physical network fabric management -- the open source way.
Introduction to the Helium release of OpenDaylightSDN Hub
"Helium" is the second release of OpenDaylight made on Oct 2, 2014. This release has more expanded support for Yang, modeling and autogeneration of REST API, improved performance of MD-SAL datastore using Tree-based Akka storage, better integration with OpenStack Neutron API, support for Group-based Policy and support for Service Function Chaining.
OpenDaylight: an open source SDN for your OpenStack cloudAnees Shaikh
Presented at the 2013 OpenStack Summit in Hong Kong.
Authors: Stephan Baucke, Kyle Mestery, Anees Shaikh, Chris Wright
OpenDaylight is an exciting new community-led, open source project focused on accelerating adoption of software-defined networking (SDN) by providing a robust SDN platform on which the industry can build and innovate. An OpenDaylight controller provides flexible management of both physical and virtual networks. The open source nature of the project and its flexible network management capabilities make it an ideal SDN platform to integrate with Neutron.
In this session, OpenDaylight community members from Cisco, IBM, RedHat, and Ericsson will describe the OpenDaylight project goals and platform architecture, as well as the roadmap and progress to date. OpenDaylight brings together a number of virtual networking approaches, and we will discuss integration approaches with OpenStack Neutron that provide flexibility for OpenStack administrators and users. Details of our initial Neutron integration will also be demonstrated for attendees.
Attendees will leave this session with a greater understanding of what OpenDaylight is, and how it can integrate with OpenStack Neutron to provide a powerful SDN-based networking solution for OpenStack Clouds.
Presentation delivered at LinuxCon China 2017.
Open vSwitch (OVS) is a multilayer open source virtual switch. OVS is designed to enable massive network automation through programmatic extension, while still supporting standard management interfaces. OVN is a new network virtualization project that brings virtual networking to the Open vSwitch user community. OVN includes logical switches and routers, security groups, and L2/L3/L4 ACLs, implemented on top of a tunnel-based overlay network.
In this presentation, we will provide an overview of the current state of the projects and their future plans, such as:
- The current state of the Linux, DPDK, and Hyper-V ports
- A status update on a portable BPF-based datapath
- The latest stateful and OpenFlow features available in OVS
- Performance and debugging enhancement to OVN
- OVN features under development such as ACL logging and encrypted tunnels
Improving Network Application Performance using Load Aware LibeventdevMichelle Holley
Compared to load unaware packet distribution mechanisms often used in the run to completion model, an event scheduler improves core utilization and better handles dynamic traffic mixes by scheduling packets to cores according to their load. It simultaneously provides both atomicity and packet ordering. Hardware-based event schedulers can also provide low-latency inter-core communication. The libeventdev library from Data Plane Development Kit (DPDK) helps developers leverage the event scheduler model.
About the presenter: Sundar Vedantham, Intel, is a Senior Technical Manager working in the Data Center Group in Allentown, PA. His research interests include network traffic and congestion management, high-speed networking, and theoretical computer models, areas in which he holds patents and has published papers, book chapter & articles. He received his Ph.D. in Computer Science in 1997 from Louisiana State University. He enjoys writing articles in English and Tamil to help improve public understanding of technical details behind the fields he has worked on and to attract young students to get into STEM fields.
VMware ESXi - Intel and Qlogic NIC throughput difference v0.6David Pasek
We are observing different network throughputs on Intel X710 NICs and QLogic FastLinQ QL41xxx NIC. ESXi hardware supports NIC hardware offloading and queueing on 10Gb, 25Gb, 40Gb and 100Gb NIC adapters. Multiple hardware queues per NIC interface (vmnic) and multiple software threads on ESXi VMkernel is depicted and documented in this paper which may or may not be the root cause of the observed problem. The key objective of this document is to clearly document and collect NIC information on two specific Network Adapters and do a comparison to find the difference or at least root cause hypothesis for further troubleshooting.
Opensource approach to design and deployment of Microservices based VNFMichelle Holley
Microservice is gaining increased adoption in the Telco NFV world. It is key to understand the design and deployment methodologies involved in developing Microservice based VNF. This talk provides an opensource practitioner approach to building and deploying a Microservice based VNF and includes the following: - Design patterns, workflow models - Design models for VNF placement, capacity management, scale-in/out and resiliency - Deployment considerations that includes handing of scale and fault tolerant VNF using well known Opensource tools.
About the presenter: Prem Sankar works for Ericsson Opensource Ecosystem team and part of the Opendaylight and OPNFV team in Ericsson. Prem evangelizes SDN and Cloud and has given many sessions and conducted workshops around SDN and ODL. Prem is PTL of ODL COE project and currently driving the Kuberenetes and ODL Integration in Opendaylight community. Prem is a frequent speaker at opensource summits and has presented in Opendaylight, OPNFV and Open networking summits.
HKG15-110: ODP Project Update
---------------------------------------------------
Speaker: Bill Fischofer
Date: February 9, 2015
---------------------------------------------------
★ Session Summary ★
This session provides a summary of ODP activities since LCU ‰Û÷14 and highlights the main features of ODP v1.0 for applications as well as the validations used by conforming ODP implementation.
--------------------------------------------------
★ Resources ★
Pathable: https://hkg15.pathable.com/meetings/250771
Video: https://www.youtube.com/watch?v=xABcGPOCOuU
Etherpad: http://pad.linaro.org/p/hkg15-110
---------------------------------------------------
★ Event Details ★
Linaro Connect Hong Kong 2015 - #HKG15
February 9-13th, 2015
Regal Airport Hotel Hong Kong Airport
---------------------------------------------------
http://www.linaro.org
http://connect.linaro.org
Looking for a way to deploy a stable OpenStack Cloud Environment with Opendaylight at ease? This session is about learning to deploy a Cloud environment with OPNFV Fuel deployer. Fuel is a deployment tool which deploys a wide variety of distributions with third party plugins like OpenDayLight, while abstracting out complexities of the deployment. The intent of this session is to familiarize deployment of OpenStack with OpenDaylight.
About the author: Pramod is a software developer in OpenStack and OpenDayLight, working for OTC, SSG at Intel. His Area of Interest is in Cloud Networking and Applications. He has prior experience in Databases and his current focus is on developing features of Cloud Networking Platform. He holds Masters Degree from San Jose State University.
June Boston openStack Summit: Preparing quantum for the data centerKamesh Pemmaraju
Quantum, OpenStack's virtual networking service, is slated to replace Nova's networking service in the upcoming Folsom release. Initial Quantum development has focused on the basic requirements of a green-field Nova cloud deployment. Additional requirements will be discussed that enable Quantum to support private Nova cloud deployments within existing data centers, as well as enterprise virtualization with the oVirt project. Initial work is underway for Folsom, and additional work will follow.
LCU14 310- Cisco ODP
---------------------------------------------------
Speaker: Robbie King
Date: September 17, 2014
---------------------------------------------------
★ Session Summary ★
Cisco to present their experience using ODP to provide portable accelerated access to crypto functions on various SoCs.
---------------------------------------------------
★ Resources ★
Zerista: http://lcu14.zerista.com/event/member/137757
Google Event: https://plus.google.com/u/0/events/ckmld1hll5jjijq11frbqmptet8
Video: https://www.youtube.com/watch?v=eFlTmslVK-Y&list=UUIVqQKxCyQLJS6xvSmfndLA
Etherpad: http://pad.linaro.org/p/lcu14-310
---------------------------------------------------
★ Event Details ★
Linaro Connect USA - #LCU14
September 15-19th, 2014
Hyatt Regency San Francisco Airport
---------------------------------------------------
http://www.linaro.org
http://connect.linaro.org
OpenStack is a free and open-source software platform for cloud computing, mostly deployed as an infrastructure-as-a-service (IaaS). OpenDaylight is an open source project under the Linux Foundation with the goal of furthering the adoption and innovation of SDN through the creation of a common industry supported platform.
In this session, I will talk about how OpenStack and OpenDaylight can be combined together to solve real world business cases and networking needs. We will cover:
- What is OpenDaylight
- Use cases for OpenDaylight with OpenStack
- The OpenDaylight NetVirt project
- How OpenDaylight interacts with OpenStack
- The future of OpenDaylight, and how we see it help solving challenges in the networking industry such as NFV, container networking and physical network fabric management -- the open source way.
Introduction to the Helium release of OpenDaylightSDN Hub
"Helium" is the second release of OpenDaylight made on Oct 2, 2014. This release has more expanded support for Yang, modeling and autogeneration of REST API, improved performance of MD-SAL datastore using Tree-based Akka storage, better integration with OpenStack Neutron API, support for Group-based Policy and support for Service Function Chaining.
LAS16-301: OpenStack on Aarch64, running in production, upstream improvements...Linaro
LAS16-301: OpenStack on Aarch64, running in production, upstream improvements, and interoperability
Speakers: Yibo Cai, Gema Gomez Solano, Jack He, Marcin Juskiewicz, Martin Stadtler
Date: September 28, 2016
★ Session Description ★
“OpenStack is at the heart of the next generation of the opensource
cloud on a global scale. During this presentation, we will touch on three themes, running an OpenStack based cloud in production by Gema Gomez and Andy Doan, followed by Marcin talking about the packaging and bug fixing on archives required to make that happen on AArch64. Jack He and Yibo Cai, will explain what it is like working with the the upstream project, the development environment, the current patches and what needs to be done next. Then Gema Gomez will Introduce the OpenStack Interop Working Group. Why is interoperability important for OpenStack? And What is Linaro doing to improve the interoperability of OpenStack?
★ Resources ★
Etherpad: pad.linaro.org/p/las16-301
Presentations & Videos: http://connect.linaro.org/resource/las16/las16-301/
★ Event Details ★
Linaro Connect Las Vegas 2016 – #LAS16
September 26-30, 2016
http://www.linaro.org
http://connect.linaro.org
HKG15-405: Redundant zero/sign-extension elimination in GCCLinaro
HKG15-405: Redundant zero/sign-extension elimination in GCC
---------------------------------------------------
Speaker: Kugan Vivekanandarajah
Date: February 12, 2015
---------------------------------------------------
★ Session Summary ★
Several instances of redundant zero/sign-extension related bugs are reported in GCC and Linaro bugzilla. These bugs are sources of performance/code size penalties. This presentation will discuss the history, design considerations, implementation, and performance characteristics of redundant zero/sign extension elimination in GCC. We will then discuss a new compiler pass that performs computation in promoted type mode in such a way that removes redundant zero/sign-extensions.
--------------------------------------------------
★ Resources ★
Pathable: https://hkg15.pathable.com/meetings/250833
Video: https://www.youtube.com/watch?v=JkTkmGe3tms
Etherpad: http://pad.linaro.org/p/hkg15-405
---------------------------------------------------
★ Event Details ★
Linaro Connect Hong Kong 2015 - #HKG15
February 9-13th, 2015
Regal Airport Hotel Hong Kong Airport
---------------------------------------------------
http://www.linaro.org
http://connect.linaro.org
AArch64 and ARM GDB ports were added some years ago, but some useful features are still missing. We started to add these features to GDB in 2015 and most of them are already accepted by the GDB mainline.
This presentation will discuss these new added features, such as reverse debugging, tracepoint, and multi-arch debugging, together with some explanations on how does GDB support them in general.
This presentation will also introduce some basic GDB or debugger internal knowledges and also some GDB in-progress projects in which we plan to do and are interested in.
ILP32 is a programming model that may be useful on AArch64 systems for performance and also for legacy code with 32-bit data size assumptions. We combined ILP32 support from upstream projects with the LEAP distribution to enable experimentation with this model. This talk discusses the relative benchmark performance of the LP64 and ILP32 programming models under AArch64.
BKK16-504 Running Linux in EL2 VirtualizationLinaro
Running Linux in EL2 offers potentially important performance benefits for running VMs at the cost of more complicated low-level code paths in the kernel and worse performance for userspace applications. This talk explores the required actions taken so far, an analysis of the benefits, and discusses challenges with upstreaming this approach.
HKG15-400: Next steps in KVM enablement on ARMLinaro
HKG15-400: Next steps in KVM enablement on ARM
---------------------------------------------------
Date: February 12, 2015
---------------------------------------------------
★ Session Summary ★
Next steps in KVM enablement on ARM
--------------------------------------------------
★ Resources ★
Pathable: https://hkg15.pathable.com/meetings/250827
Video: https://www.youtube.com/watch?v=g8noeSpWVDY
Etherpad: http://pad.linaro.org/p/hkg15-400
---------------------------------------------------
★ Event Details ★
Linaro Connect Hong Kong 2015 - #HKG15
February 9-13th, 2015
Regal Airport Hotel Hong Kong Airport
---------------------------------------------------
http://www.linaro.org
http://connect.linaro.org
LAS16-500: The Rise and Fall of Assembler and the VGIC from HellLinaro
LAS16-500: The Rise and Fall of Assembler and the VGIC from Hell
Speakers: Marc Zyngier, Christoffer Dall
Date: September 30, 2016
★ Session Description ★
KVM/ARM has grown up. While the initial implementation of virtualization support for ARM processors in Linux was a quality upstream software project, there were initial design decisions simply not suitable for a long-term maintained hypervisor code base. For example, the way KVM/ARM utilized the hardware support for virtualization, was by running a ‘switching’ layer of code in EL2, purely written in assembly. This was a reasonable design decision in the initial implementation, as the switching layer only had to do one thing: Switch between a VM and the host. But as we began to optimize the implementation, add support for ARMv8.1 and VHE, and added features such as debugging support, we had to move to a more integrated approach, writing the switching logic in C code as well. As another example, the support for virtual interrupts, famously known as the VGIC, was designed with a focus on optimizing MMIO operations. As it turns out, MMIO operations is a less important and infrequent operation on the GIC, and the design had some serious negative consequences for supporting other state transitions for virtual interrupts and had negative performance implications. Therefore, we completely redesigned the VGIC support, and implemented the whole thing from scratch as a team effort, with a very promising result, upstream since Linux v4.7. In this talk we will cover the evolution of this software project and give an overview of the state of the project as it is today.
★ Resources ★
Etherpad: pad.linaro.org/p/las16-500
Presentations & Videos: http://connect.linaro.org/resource/las16/las16-500/
★ Event Details ★
Linaro Connect Las Vegas 2016 – #LAS16
September 26-30, 2016
http://www.linaro.org
http://connect.linaro.org
SFO15-407: Performance Overhead of ARM VirtualizationLinaro
SFO15-407: Performance Overhead of ARM Virtualization
Speaker: Christoffer Dall
Date: September 24, 2015
★ Session Description ★
The ARM architecture’s support for virtualization has been designed for Type I hypervisors such as Xen. However, Type II hypervisors, such as KVM, offer increased convenience, portability, and an integrated software development environment with existing projects such as the Linux kernel and QEMU. Since the architecture has been designed with a separate CPU mode for the hypervisor, Type II hypervisors must perform the notorious "double-trap" to switch between the guest and the host, and many people have speculated that this causes KVM to be inherently slower than for example Xen. The truth is as always much more complicated than that, and it turns out that the differences in performance between Xen and KVM on the ARMv8 architecture come from entirely different sources. This talk will present a number of performance figures from running KVM and Xen on ARMv8 server hardware and will compare these numbers to x86 servers to illustrate the viability of virtualization on ARMv8 for the enterprise and networking markets.
★ Resources ★
Video: N/A
Presentation: http://www.slideshare.net/linaroorg/sfo15407-performance-overhead-of-arm-virtualization
Etherpad: pad.linaro.org/p/sfo15-407
Pathable: https://sfo15.pathable.com/meetings/303079
★ Event Details ★
Linaro Connect San Francisco 2015 - #SFO15
September 21-25, 2015
Hyatt Regency Hotel
http://www.linaro.org
http://connect.linaro.org
LAS16-200: SCMI - System Management and Control InterfaceLinaro
Title: SCMI - System Management and Control Interface
Abstract: In this session we present a new standard proposal for system control and management. The industry, both in high end mobile and enterprise, is trending towards the use of power and system controllers. In most cases the controllers have very similar communication mechanisms between application processors and controllers. In addition, these controllers generally provide very similar functions, e.g. DVFS, power domain management, sensor management. This standard proposal provides an extensible, OS agnostic, and virtualizable interface to access these functions.
Speaker(s):Charles Garcia-Tobin
USENIX LISA15: How TubeMogul Handles over One Trillion HTTP Requests a MonthNicolas Brousse
TubeMogul grew from few servers to over two thousands servers and handling over one trillion http requests a month, processed in less than 50ms each. To keep up with the fast growth, the SRE team had to implement an efficient Continuous Delivery infrastructure that allowed to do over 10,000 puppet deployment and 8,500 application deployment in 2014. In this presentation, we will cover the nuts and bolts of the TubeMogul operations engineering team and how they overcome challenges.
Workday has built one of the largest OpenStack-based private clouds in the world, hosting a workload of over a million physical cores on over 16,000 compute nodes in 5 data centers for over ten years. However, there was a growing need for a newer, more maintainable deployment model that would closely follow the upstream community. We would like to share our new architecture and deployment approach as well as lessons learned from our experience.
We’ve converted many of our technologies in the process, from…
Migrating from Mitaka, to Victoria
Converting from OpenContrail, to pure L3 Calico with BGP on the host
Deploying with Chef, to deploying with Ansible
Building home-grown container images, to Kolla
Monitoring with Sensu and Wavefront, to Prometheus and Grafana
CI/CD in Jenkins, to Zuul
CentOS 7, to CentOS 8 Stream
We'll also talk about some internal tools we wrote that, while Workday-specific, may inspire you to see what value-add you can make for your customers.
Xpdays: Kubernetes CI-CD Frameworks Case StudyDenys Vasyliev
A set of flexible and comprehensive operation principles to cover all stages of a modern application life cycle.
Almost any Customer wants the Setup to be compatible with existing infrastructure. It assumes a Bare Metal, Private or Public Cloud. In special cases even offline setup, for example, Airports, Fintech sector or Telecom operators. The main requirements are: Scalability, High Availability, Security Compliance, Professional Service.
So, we should cover all three tiers: Infrastructure, Control Plane and Application Plane. Market leaders are Drone, Argo and Knative. And our story we called Cloud Flex Framework.
Lisa Caywood and Colin Dixon's presentation at the 2017 Open Networking Summit.
OpenDaylight has become a nexus for open source integration, creating a new open networking stack and enabling a new generation of open source, agile IT infrastructure. The fifth “Boron” release provides new tooling and documentation to support application developers, as well as greater integration with industry frameworks from OPNFV and OpenStack to CORD and Atrium. Boron also brings a practical focus on two leading types of deployments: (1) direct control of virtual switches to provide network virtualization and NFV and (2) management and orchestration of existing networks to provide new features and automation. This talk will cover trends in open SDN and cloud networking, with a focus on Boron milestones. In particular, it dives into the architecture across OpenStack and OpenDaylight to enable OpenStack service function chaining support in OpenDaylight.
In this deck, Paul Isaacs from Linaro presents: State of ARM-based HPC. This talk provides an overview of applications and infrastructure services successfully ported to Aarch64 and benefiting from scale.
"With its debut on the TOP500, the 125,000-core Astra supercomputer at New Mexico’s Sandia Labs uses Cavium ThunderX2 chips to mark Arm’s entry into the petascale world. In Japan, the Fujitsu A64FX Arm-based CPU in the pending Fugaku supercomputer has been optimized to achieve high-level, real-world application performance, anticipating up to one hundred times the application execution performance of the K computer. K was the first computer to top 10 petaflops in 2011."
Watch the video: https://wp.me/p3RLHQ-lIT
Learn more: https://www.linaro.org/
Sign up for our insideHPC Newsletter: http://insidehpc.com/newsletter
Webinar: OpenEBS - Still Free and now FASTEST Kubernetes storageMayaData Inc
Webinar Session - https://youtu.be/_5MfGMf8PG4
In this webinar, we share how the Container Attached Storage pattern makes performance tuning more tractable, by giving each workload its own storage system, thereby decreasing the variables needed to understand and tune performance.
We then introduce MayaStor, a breakthrough in the use of containers and Kubernetes as a data plane. MayaStor is the first containerized data engine available that delivers near the theoretical maximum performance of underlying systems. MayaStor performance scales with the underlying hardware and has been shown, for example, to deliver in excess of 10 million IOPS in a particular environment.
Como creamos QuestDB Cloud, un SaaS basado en Kubernetes alrededor de QuestDB...javier ramirez
QuestDB es una base de datos open source de alto rendimiento. Mucha gente nos comentaba que les gustaría usarla como servicio, sin tener que gestionar las máquinas. Así que nos pusimos manos a la obra para desarrollar una solución que nos permitiese lanzar instancias de QuestDB con provisionado, monitorización, seguridad o actualizaciones totalmente gestionadas.
Unos cuantos clusters de Kubernetes más tarde, conseguimos lanzar nuestra oferta de QuestDB Cloud. Esta charla es la historia de cómo llegamos ahí. Hablaré de herramientas como Calico, Karpenter, CoreDNS, Telegraf, Prometheus, Loki o Grafana, pero también de retos como autenticación, facturación, multi-nube, o de a qué tienes que decir que no para poder sobrevivir en la nube.
OSDC 2018 | Highly Available Cloud Foundry on Kubernetes by Cornelius SchumacherNETWAYS
This presentation will show how Cloud Foundry, the popular Platform as a Service framework, is deployed and configured to run in a highly available fashion on Kubernetes. It will show how to avoid single points of failures using Kubernetes features like stateful sets, readiness and liveness probes, etc. This includes how high availability extends to applications deployed by the end users of Cloud Foundry so they don’t have to worry about downtime. The presentation will include a demo of a disruptive agent simulating failures across the Kubernetes nodes and containers, while user applications are still alive and healthy. This presentation shows a real-life production use case for Kubernetes. This can be used as an example and to learn about the high-availability related features of Kubernetes. It also presents how the Kubernetes stack can be extended with Cloud Foundry to also cover the use case of Platform as a Service.
Using Kubernetes to make cellular data plans cheaper for 50M usersMirantis
Use case of Kubernetes based NFV infrastructure used in production to run an open source evolved packet core. Presented by Facebook Connectivity and Mirantis at KubeCon + CloudNativeCon Europe 2020.
2. ARM Progress in Servers
ARM-based servers on track
▪ 16 licenses now signed for server applications
▪ 2013: Chip development
▪ 2014: Software optimization accelerates
▪ 2015: First production systems
▪ 2018: ARM expects 10-15% share
of total server market
Data Center Workloads are Changing
Today Next 3 Years 5 Years +
Large
discrete
tasks
3. Linaro Overview
● Linaro is a collaborative software
engineering company funded by
members
● Instead of duplicating effort for
common open source software,
competitors share development
costs and software is built once
● The work is carried out in the open,
tested and then upstreamed into the
relevant open source projects:
kernel.org, Gnu, Fedora, Ubuntu,
OpenJDK etc.
4. Linaro Organization
● Founded 2010
● Now 30 members
● $50M annual revenue
● >200 OSS engineers
● Global footprint
Technical Steering Committee (TSC)
Office of the CTO (OCTO)
Enterprise Group (LEG)
16 companies
Platform Engineering
Builds & Baselines
QA Services
Systems
LAVA Lab
Field Engineering
Digital Home Group (LHG)
8 companies
Mobile Group (LMG)
10 companies
Networking Group (LNG)
12 companies
Core Development
Kernel
Power Management
Security
Virtualization
Product Technology
LAVA Software Linaro Stable Kernel (LSK) Toolchain
Member
Services
Club/Core
Services
Linaro Developer
Technical Support
(LDTS)
6. Linux Kernel Upstreaming
● Linux 3.17 October 2014
● 12,353 total patch sets
in this release
● 180 known companies
contributed
● Linaro is #4 company
contributor; in top 5 for last
8 kernel versions (since 3.10)
Top Linux 3.17 Patchset Contributors by Company
1 Unknown/Hobbyists 3705 (30.00%)
2 Intel 1261 (10.21%)
3 Red Hat 980 (7.93%)
4 Samsung 491 (3.97%)
5 Linaro 412 (3.34%)
6 Vision Engraving & Routing 393 (3.18%)
7 Novell 326 (2.64%)
8 Texas Instruments 306 (2.48%)
9 Renesas Electronics 263 (2.13%)
10 Google 227 (1.84%)
Source: http://www.remword.com/kps_result/3.17_whole.html
8. Linaro Segment Groups
Digital Home - LHG Mobile - LMG Networking - LNG Enterprise - LEG
Confidential
● OSS for the digital home
● W3C EME Secure
Media playback for
RDK and Android
● Middleware and user-space
stack
● DRM, DLNA, CVP-2,
HTML5
● LSK kernel version for
STB/IPTV
● Common media
frameworks
● OSS for mobile devices
● Android 64 bit, “L”
● big.LITTLE power
management
● QEMU based Android
development
● 64-bit Chromium browser
● Support members,
ARM and Google
Android development
● OSS for networking
● Real Time Support
● Virtualization
● Core isolation
● OpenDataPlane (ODP)
● Big-endian legacy support
● ODP cross-platform
support for SoC
accelerators
● OSS for ARM servers
● UEFI/ACPI
● KVM/Xen
● ARMv8 optimization
● OpenJDK, Hadoop,
OpenStack
● Reduces fragmentation,
cost, accelerates time to
market
9. Linaro Enterprise Group (LEG)
● Formed in November 2012
● enable the ARM server ecosystem
● share costs of the enablement
engineering
● focus on standards, avoid
fragmentation
● identify, test and optimise
workloads and core libraries
http://wiki.linaro.org/LEG
10. Engineering (1): enablement
● All patches already upstream or under deep constructive review
● Strong engagement with all maintainers
12. Engineering I
● LEG projects agreed and prioritized by LEG Steering Committee
● One representative per member, plus Linaro LEG Director
● ARM Boot Architecture
● ARM Server Base System Architecture; ACPI, UEFI, Grub; ARM Trusted firmware
● KVM and Xen Virtualization
● Stable ARM kernel version based on upstream
● LAMP Stack optimization
● HDFS CRC, LibTBB, CRC, Hugepages, OpenSSL
● Facebook HipHop JIT
● High performance OpenJDK for ARMv8 with C1 and C2 JIT
13. Engineering II
● Middleware and user-space stack testing
● Hadoop, OpenStack, Ceph
● Testing through multinode CI validation in LAVA
● ARMv8 Ubuntu, Fedora/Red Hat and OpenStack builds
● LEG also works with Linaro core working groups and can request
resources through the Linaro Technical Steering Committee (TSC)
● ARMv8 64 bit toolchain optimization
● Multi-core power management
● Security and Secure boot using Trustzone and
ARM Trusted Firmware
14. Achievements (1)
● UEFI
● Aligned ARM Tianocore to other architectures, boot as an EFI application, runtime
services, pass ACPI tables from firmware, support for SMBIOS 3.0
● Support GRUB and network boot
● Ported GNUEFI to build EFI apps easily
● Verified Tianocore Secure boot and ported the Shim layer to ARM
● Ported UEFI to EL1 as a guest in KVM/QEMU hypervisor (ongoing for Xen)
● Established a monthly UEFI release rebased on Tianocore EDK2 + all Linaro patches,
tested with UEFI SCT suite
● ACPI
● Ported ACPICA core kernel support and tools to ARM
● Ported key ARM core peripherals to ACPI
● Able to boot ARM FVP, Juno, AMD Seattle and Cavium Thunder-X model via ACPI
● Developed a PCC driver to support the new CPPC power management model
● Ported all reference test suites to ARM, e.g. FWTS, ACPI API test, ACPI ASL test, etc.
● Releasing an updated LEG ACPI kernel for every new weekly kernel 3.xx-rc
15. Achievements (2)
● OpenJDK
● Excellent cooperation with Red Hat on OpenJDK C1 and C2 JIT
● Driven by OpenJDK8, backported to OpenJDK7 and being forward ported to
OpenJDK9
● On par with x86 as JIT performance vs interpreted code
● Close to 10,000 tests executed every night: Mauve, JTREG, JCStress, SPECjbb2013
● Hadoop TeraSort and Jenkins as functional testing workloads
● Optimisations for the LAMP Webserver workload
● ARMv8 assembly tuning for OpenSSL, CRC32, etc. up to 5x-16x speed up
● Hugepages, fast_gup, etc.
● OpenStack
● Built OpenStack on ARMv8 with KVM/QEMU (ongoing for Xen)
● Running the Tempest test suite and the Rally benchmarking suite
● Functional test deploying multiple VM’s to run Java JTREG testing in parallel
● Early proof of concept with Docker and Go
17. Platform as a Service (PaaS)
Software as a Service (SaaS)
Infrastructure as a Service (IaaS)
18. Mobility as a Service (MaaS)
Mobile Backend as a service (MBaaS)
Database as a Service (DBaaS)
Logging as a Service (LaaS)
Location as a service (LaaS)
Platform as a Service (PaaS)
Payments as a service (PaaS)
Software as a Service (SaaS)
Data as a Service (DaaS) Desktop as a Service (DaaS)
Infrastructure as a Service (IaaS)
Network as a service (NaaS) IT as a service (ITaaS)
Security as a service (SECaaS)
Storage as a service (STaaS)
Recovery as a service (RaaS)
32. It’s running on ARM 64-bit now!
http://www.ca.com/us/lpg/ca-technology-exchange/sunny-outlook-for-openstack-clouds.aspx
33. It’s running on ARM 64-bit now!
● Long cooperation with Canonical and Red Hat engineers
● Reusing pre-built DevStack packages from Ubuntu
● Leveraging on LEG kernel and Linaro KVM work
● Running on real ARM 64-bit hardware,
APM X-Gene and HP Moonshot
● Running all applicable Tempest tests
● Ongoing
● PackStack packages from Fedora
● libvirt running on XEN
34. OpenJDK on ARMv8
● AARCH64 port of OpenJDK
● Based on same codebase as x86 OpenJDK
● 95%+ code in common with Oracle JDK
● Same relative C2 vs interpreter JIT performance within a few % as Oracle JDK
● Goal is to track all x86 optimisations / features
● No support for ARMv7 or earlier or for 32 bit on ARMv8
● Three releases are supported: JDK7, JDK8 and JDK9
35. CI and nightly tests
Server results
from Aug 22
2014
ARM64 x86
Pass Fail Error Pass Fail Error
Hotspot 562 2 0 562 2 0
langtools 3004 0 11 2979 0 25
JDK 5283 181 46 5273 232 5
Totals 8849 183 57 8814 234 30
36. Hadoop Testing / Optimisation
● Daily native build of Hadoop from source
● Daily benchmark testing
● Client and server run of Terasort to test for performance regression
● Single node testing only.
● Looking at multinode testing
● Optimisation focused on optimising OpenJDK for Hadoop
● +24% improvement since FCS release in April
● Some optimisation on Hadoop directly
● CRC optimisation using SIMD
● Pushed to list but not yet adopted upstream
38. Demo Overview
● 3x APM Mustang systems running Ubuntu Trusty 14.04
● All services deployed with juju
● Mustang 01 as cloud controller and jenkins master with LXC
● Mustang 02 and 03 as OpenStack compute nodes with KVM
● Run a subset of the OpenJDK JTREG test regression harness
● jenkins dynamically provisions SIX VM’s instances via OpenStack on Mustang 2 and 3
● the VM’s are destroyed upon test completion
39. What’s the effort and cost to port OpenStack?
let’s review few open source projects
let’s rely on ohloh.net and the co.co.mo. model
http://csse.usc.edu/csse/research/COCOMOII/cocomo_main.html
48. Comparison
Project SLOC (M) Commits/month Start date
OpenStack 2 3992 2010*
OpenJDK 8 5 983 2007
Hadoop 2.5 219 2006
mongoDB 0.6 500 2007
Linux Kernel 17 6000 2002
Google Chrome 8 5000 2008
(*) 1st commit for OpenStack was in 2006 but really picked up in mid 2010
● OpenStack is still relatively small but growing as fast as the Linux
Kernel itself…
50. The Fonzarelli Fix
a.k.a. percussive maintenance
Tapping the server with a hammer: $0.00.
Knowing where to hit it: $5,000.00.
http://tvtropes.org/pmwiki/pmwiki.php/Main/PercussiveMaintenance
51. Patches available upstream
● nova: "Support setting a machine type to enable ARMv7/AArch64 guests to boot"
http://git.openstack.org/cgit/openstack/nova/commit/?id=15d4f8a3af1b4e105d91e6b66d7b19187fbd41af
● libvirt: "AArch64: Add AArch64 architecture to list of valid arches."
http://libvirt.org/git/?p=libvirt.git;a=commit;h=670c08afd461cba6164f7a8a643f5a5b41782e8d
● libvirt: "AArch64: Porting of armv7l conditons to run qemu for aarch64."
http://libvirt.org/git/?p=libvirt.git;a=commit;h=27e32e0f3d4d5840cee7e361bdff94c68ae181e2
● libvirt: "AArch64: CPU Support for AArch64 (ARMv8 64bit)."
http://libvirt.org/git/?p=libvirt.git;a=commit;h=5cb6816715968dc08486307bd1894dc91e22342c
● libvirt: "qemu: Add support for virt machine type with virtio-mmio devices on armv7"
http://libvirt.org/git/?p=libvirt.git;a=commit;h=c7ccd2c44be99b81ffce29374faeab3d1644a151
52. LibVirt patches
● nova: "Support setting a machine type to enable ARMv7/AArch64 guests to boot"
http://git.openstack.org/cgit/openstack/nova/commit/?id=15d4f8a3af1b4e105d91e6b66d7b19187fbd41af
● libvirt: "AArch64: Add AArch64 architecture to list of valid arches."
http://libvirt.org/git/?p=libvirt.git;a=commit;h=670c08afd461cba6164f7a8a643f5a5b41782e8d
● libvirt: "AArch64: Porting of armv7l conditons to run qemu for aarch64."
http://libvirt.org/git/?p=libvirt.git;a=commit;h=27e32e0f3d4d5840cee7e361bdff94c68ae181e2
● libvirt: "AArch64: CPU Support for AArch64 (ARMv8 64bit)."
http://libvirt.org/git/?p=libvirt.git;a=commit;h=5cb6816715968dc08486307bd1894dc91e22342c
● libvirt: "qemu: Add support for virt machine type with virtio-mmio devices on armv7"
http://libvirt.org/git/?p=libvirt.git;a=commit;h=c7ccd2c44be99b81ffce29374faeab3d1644a151
Thanks to libvirt maintainer Cole Robinson at Red Hat
for his support!
53. OpenStack patches: only 114 lines in Nova
● nova: "Support setting a machine type to enable ARMv7/AArch64 guests to boot"
http://git.openstack.org/cgit/openstack/nova/commit/?id=15d4f8a3af1b4e105d91e6b66d7b19187fbd41af
Thanks to Canonical engineers
for their contribution!
54. Deployment
Ready
Software
Developer ARMv8 Server Cluster to Accelerate
Deployment-ready Software
Operating
System
Vendors
+ +
Open
Source
Community
Independent
Software
Vendors
www.linaro.org/leg/servercluster/
55. Summary
● OpenStack is at the heart of the cloud revolution
● All workloads run well on Linaro ARMv8 foundations
● Linaro fosters the open source collaboration on ARM
56. More about Linaro: www.linaro.org/about/
Linaro members: www.linaro.org/members