DOCSLIB.ORG

Trusted Docker Containers and Trusted Vms in Openstack

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Trusted Docker Containers and Trusted Vms in Openstack Trusted Docker Containers and Trusted VMs in OpenStack Raghu Yeluri Abhishek Gupta Outline o Context: Docker Security – Top Customer Asks o Intel’s Focus: Trusted Docker Containers o Who Verifies Trust ? o Reference Architecture with OpenStack o Demo o Availability o Call to Action Docker Overview in a Slide.. Docker Hub Lightweight, open source engine for creating, deploying containers Provides work flow for running, building and containerizing apps. Separates apps from where they run.; Enables Micro-services; scale by composition. Underlying building blocks: Linux kernel's namespaces (isolation) + cgroups (resource control) + .. Components of Docker Docker Engine – Runtime for running, building Docker containers. Docker Repositories(Hub) - SaaS service for sharing/managing images Docker Images (layers) Images hold Apps. Shareable snapshot of software. Container is a running instance of image. Orchestration: OpenStack, Docker Swarm, Kubernetes, Mesos, Fleet, Project Docker Layers Atomic, Lattice… Docker Security – 5 key Customer Asks 1. How do you know that the Docker Host Integrity is there? o Do you trust the Docker daemon? o Do you trust the Docker host has booted with Integrity? 2. How do you verify Docker Container Integrity o Who wrote the Docker image? Do you trust the image? Did the right Image get launched? 3. Runtime Protection of Docker Engine & Enhanced Isolation o How can Intel help with runtime Integrity? 4. Enterprise Security Features – Compliance, Manageability, Identity authentication.. Etc. 5. OpenStack as a single Control Plane for Trusted VMs and Trusted Docker Containers.. Intel’s Focus: Enable Hardware-based Integrity Assurance for Docker Containers – Trusted Docker Containers Trusted Docker Containers – 3 focus areas o Launch Integrity of Docker Host o Runtime Integrity of Docker Host o Integrity of Docker Images Today’s Focus: Integrity of Docker Host, and how to use it in OpenStack. Trusted VMs - Summary o Launch VMs on Servers that have demonstrated Boot Integrity – Platform Trust Apps Measurements done at the time of Server boot) o Measured Launch of Boot Process/Components vFW VM-2 with Intel TXT. Host OS/Hypervisor o Trust Chain: HW->FW->BIOS->OS/VMM Kernel, Initrd o Tboot What is measured at launch: HW w/ Intel TXT/TPM o Current: F/W, Core BIOS, OS/VMM Kernel, Initrd Measurements match! System trusted o Ext measurements: An7OS/FS modules Trust Boundary o Schedulers/Orchestrators Policy Manager use Trust to launch/create/Migrate VMs. App App o Measurements done at the time of boot Extend Chain of Trust to VMs. (Server boot and VM Launch) o Measure & Attest VM Images prior to Launch. VM-1 VM-2 Trust Boundary o Encrypt VM Images and decrypt based on Platform Trust (Tenant-Controls the Keys) vRTM Host OS/Hypervisor Kernel, Initrd++ o Boundary Control of VMs– Control where your Tboot Trusted VMs are launching and migrating. HW w/ Intel TXT/TPM Measurements match! System & VMs Trusted Will enable the same model and use-cases for Trusted Docker Containers Trusted Docker Containers - 1 Ensure Docker Containers are launched on Trusted Docker Hosts Container B container A Container C e.g. e.g. Apache o Boot-time integrity of the Docker Host e.g. Nginx Apache v2 Shared Bin/Libs o Measured Launch of Boot Process & components with Intel TXT. Docker Daemon o Docker daemon and associated Host OS component added to TCB and Measured. TBOOT TPM HW w/ Intel TXT o Chain of Trust: H/w->FW->BIOS->OS- >Docker Engine Docker Host Platform Integrity o Remote attestation using an Attestation Authority* Trusted Docker Containers - 2 Ensure that Docker Images are not tampered } container Container prior to Launch - Container A B o Launch time integrity of Docker Images C e.g. e.g. e.g. Nginx Apache Apache v2 o Chain of Trust: H/w->FW->BIOS->OS->Docker Engine -> Docker container Shared Bin/Libs layers (apache, Ubuntu14.04, ubuntu14,…, base) o Docker daemon modification: prior to container launch, measure and verify Agents Docker Daemon Docker image (and parent layer graph recursively) Host OS Boundary Control/Geo-Tagging applies equally TBOOT to Docker Containers as well - Compliance TPM HW w/ Intel TXT Needs. Docker Host & Container Launch o Orchestrator determines location/boundary for launching Integrity Docker Images. Exploring: Docker Image encryption & Trust- based Retrieval of Keys – Sensitive Container Images (VNFs, PCI-DSS/HIPPA Containers.. etc) What is measured for Trusted Docker Containers Apache Patch v2 Trusted launch of containerized application Apache Patch v1 Docker Daemon Apache • container management engine (e.g. Docker engine) Ubuntu14.04 • Measurement Agents Ubuntu Containerized application Initrd++ (includes a measurement agent) layers (e.g. Docker image layers) Bootloader, Tboot and OS Kernel launch Bios ChainofTrustextendedtoapplication ACM signed by manufacturer Intel® TXT + TPM What is measured – the details PCR0 + SINIT System Hash + … Power ON PCR0 PCR0 PCR0+ ENTERACCS: PCR17 PCR18 PCR19 PCR19+ LockConfig SENTER ACM Measure UCode Init TXT Lock SINIT Tboot-xm Validates, SMM & Non- Load uCode SINIT Validates, & Mem, TXT & Measures Measures Measures other Critical SINIT & Validates Measures Launch Measures Load Memory OS Kernel Docker BIOS Init Trusted Code OS code SINIT TBOOT OS BIOS ACM SMM Config Initrd++ Engine, other Code Code BIOS OS Option ROMs & other X non-critical modules Measurement Phase 1 (H/W + BIOS) Measurement Phase II (TBOOT, OS, Docker • uCode evals BIOS ACM Engine…) • BIOS ACM (evals BIOS init code) • Boot loader • BIOS • uCode (evals SINIT ACM) • BIOS Option ROMs • SINIT ACM (measures OS Kernel, initrd •Tboot-xm(agent in initrd) measures DockerEngine, other components Source: Intel Who Verifies the Docker Host Trust? Scheduler/Cluster Manager/Policy Manager… Image Docker Principles Of Operation Agents Registry Engine OS, Initrd++ o Cluster Manager determines best TPM v1.2 hosts in the cluster, based on Trusted utilization, type, location Scheduler/Cluster Manager Host compliance.. etc. Trust Filter Docker o (for this host list) Cluster Manager Agents Engine verifies Host Integrity with the OS/initrd+ Examples Attestation Authority. TPM v1.2 Attestation • OpenStack Traffic o Attestation Authority responds with • Docker Trusted Swarm Host Attestation Reports for the Hosts • Kubernetes Remote Docker o • Attestation Agents Cluster Manager picks best Server Mesos API Engine • Fleet Attestation OS,Initrd+ that has the Integrity and Authority instantiates Containers. TPM v1.2 Trust Not Verified. Trusted Docker Containers & VMs with OpenStack 5 Nova + Docker Nova Docker Agents +Agents Engine Engine Glance OS, Initrd++ OS, Initrd++ Nova Scheduler TPM v1.2 TPM v1.2 2 Trusted Trusted API Server ImageProp Filter Host Host Location Trust 3 Filter Filter 1 4 VM1 VM2 5 Nova + Nova + Remote Qemu Agents Qemu Trust VM launch Attestation Agents API Attestation OS OS Horizon Trusted Container Authority TPMTPM v1.2 v1.2 TPMTPM v1.2v1.2 Launch (OAT) Trusted Trust Not Verified. Host Horizon/API Server : Initiate Launch of Image (with 1 4 Attestation Authority: Challenges Host to Attest. Hypervisor_Type Property) Provides Signed Attestation Report to Scheduler to use. – Identifies Trusted Host for VMs or Docker Containers. 2 Nova Scheduler: ImageProp Filter excludes Hosts that don’t met Image Hypervisor Type. Nova Compute: Download Glance Image and Launch. 5 For Docker Images: Nova uses DockerDriver to download, and loaded 3 Nova Scheduler: Runs Trust/Location Filter to identify Trusted Host (for VM or Docker Container) to Docker File system with Docker load Command. [ Changes needed in OpenStack Infrastructure OpenStack changes Docker Specific changes 1. Add hypervisor_type property to images For Docker Image Integrity: Value=qemu for VM images o Modified Docker daemon to intercept Value=docker for docker images container launch request and call 2. Activate ImageProperties filter measurement agent before launch filters out hosts that don’t match Value from Image o Hypervisor Type Manifest/trust-policy created and associated with each Docker layer 3. Activate Trust filter in openstack scheduler and trust properties in images 4. Configure Nova-compute to use docker driver. DEFAULT] compute_driver = Infrastructure related changes novadocker.virt.docker.DockerDriver o TXT/TPM hardware; Steps at: https://wiki.openstack.org/wiki/Docker) o TXT/TPM activation on the clusters o Attestation Server is setup Demo Summary & Call to Action o Intel’s focus: Enable Hardware-based Integrity assurance for Docker Containers – Trusted Docker Containers o Enabling the same model as we have done for VMs. o Intel TXT and Attestation Software becomes the foundation for asserting Docker Host Integrity.. o Intel iKGT (Kernel Guard Technology) can help in runtime integrity protection of the Linux Kernel. o OpenStack can launch VMs and Containers with the extensions that are already mainstream (Trusted Compute Pools) o Get engaged, get started with Trusted VMs and OpenStack. Extensions to OpenStack for Trusted Docker containers, will be available in Q3 timeframe. o iKGT is available now on 01.org. Download it and try it out. Q & A.
Load more

Recommended publications
  • Providing User Security Guarantees in Public Infrastructure Clouds
    1 Providing User Security Guarantees in Public Infrastructure Clouds Nicolae Paladi, Christian Gehrmann, and Antonis Michalas Abstract—The infrastructure cloud (IaaS) service model offers improved resource flexibility and availability, where tenants – insulated from the minutiae of hardware maintenance – rent computing resources to deploy and operate complex systems. Large-scale services running on IaaS platforms demonstrate the viability of this model; nevertheless, many organizations operating on sensitive data avoid migrating operations to IaaS platforms due to security concerns. In this paper, we describe a framework for data and operation security in IaaS, consisting of protocols for a trusted launch of virtual machines and domain-based storage protection. We continue with an extensive theoretical analysis with proofs about protocol resistance against attacks in the defined threat model. The protocols allow trust to be established by remotely attesting host platform configuration prior to launching guest virtual machines and ensure confidentiality of data in remote storage, with encryption keys maintained outside of the IaaS domain. Presented experimental results demonstrate the validity and efficiency of the proposed protocols. The framework prototype was implemented on a test bed operating a public electronic health record system, showing that the proposed protocols can be integrated into existing cloud environments. Index Terms—Security; Cloud Computing; Storage Protection; Trusted Computing F 1 INTRODUCTION host level. While support data encryption at rest is offered by several cloud providers and can be configured by tenants Cloud computing has progressed from a bold vision to mas- in their VM instances, functionality and migration capabil- sive deployments in various application domains. However, ities of such solutions are severely restricted.
    [Show full text]
  • NOVA: a Log-Structured File System for Hybrid Volatile/Non
    NOVA: A Log-structured File System for Hybrid Volatile/Non-volatile Main Memories Jian Xu and Steven Swanson, University of California, San Diego https://www.usenix.org/conference/fast16/technical-sessions/presentation/xu This paper is included in the Proceedings of the 14th USENIX Conference on File and Storage Technologies (FAST ’16). February 22–25, 2016 • Santa Clara, CA, USA ISBN 978-1-931971-28-7 Open access to the Proceedings of the 14th USENIX Conference on File and Storage Technologies is sponsored by USENIX NOVA: A Log-structured File System for Hybrid Volatile/Non-volatile Main Memories Jian Xu Steven Swanson University of California, San Diego Abstract Hybrid DRAM/NVMM storage systems present a host of opportunities and challenges for system designers. These sys- Fast non-volatile memories (NVMs) will soon appear on tems need to minimize software overhead if they are to fully the processor memory bus alongside DRAM. The result- exploit NVMM’s high performance and efficiently support ing hybrid memory systems will provide software with sub- more flexible access patterns, and at the same time they must microsecond, high-bandwidth access to persistent data, but provide the strong consistency guarantees that applications managing, accessing, and maintaining consistency for data require and respect the limitations of emerging memories stored in NVM raises a host of challenges. Existing file sys- (e.g., limited program cycles). tems built for spinning or solid-state disks introduce software Conventional file systems are not suitable for hybrid mem- overheads that would obscure the performance that NVMs ory systems because they are built for the performance char- should provide, but proposed file systems for NVMs either in- acteristics of disks (spinning or solid state) and rely on disks’ cur similar overheads or fail to provide the strong consistency consistency guarantees (e.g., that sector updates are atomic) guarantees that applications require.
    [Show full text]
  • Firecracker: Lightweight Virtualization for Serverless Applications
    Firecracker: Lightweight Virtualization for Serverless Applications Alexandru Agache, Marc Brooker, Andreea Florescu, Alexandra Iordache, Anthony Liguori, Rolf Neugebauer, Phil Piwonka, and Diana-Maria Popa, Amazon Web Services https://www.usenix.org/conference/nsdi20/presentation/agache This paper is included in the Proceedings of the 17th USENIX Symposium on Networked Systems Design and Implementation (NSDI ’20) February 25–27, 2020 • Santa Clara, CA, USA 978-1-939133-13-7 Open access to the Proceedings of the 17th USENIX Symposium on Networked Systems Design and Implementation (NSDI ’20) is sponsored by Firecracker: Lightweight Virtualization for Serverless Applications Alexandru Agache Marc Brooker Andreea Florescu Amazon Web Services Amazon Web Services Amazon Web Services Alexandra Iordache Anthony Liguori Rolf Neugebauer Amazon Web Services Amazon Web Services Amazon Web Services Phil Piwonka Diana-Maria Popa Amazon Web Services Amazon Web Services Abstract vantage over traditional server provisioning processes: mul- titenancy allows servers to be shared across a large num- Serverless containers and functions are widely used for de- ber of workloads, and the ability to provision new func- ploying and managing software in the cloud. Their popularity tions and containers in milliseconds allows capacity to be is due to reduced cost of operations, improved utilization of switched between workloads quickly as demand changes. hardware, and faster scaling than traditional deployment meth- Serverless is also attracting the attention of the research com- ods. The economics and scale of serverless applications de- munity [21,26,27,44,47], including work on scaling out video mand that workloads from multiple customers run on the same encoding [13], linear algebra [20, 53] and parallel compila- hardware with minimal overhead, while preserving strong se- tion [12].
    [Show full text]
  • KVM/ARM: Experiences Building the Linux ARM Hypervisor
    KVM/ARM: Experiences Building the Linux ARM Hypervisor Christoffer Dall and Jason Nieh fcdall, [email protected] Department of Computer Science, Columbia University Technical Report CUCS-010-13 April 2013 Abstract ization. To address this problem, ARM has introduced hardware virtualization extensions in the newest ARM As ARM CPUs become increasingly common in mo- CPU architectures. ARM has benefited from the hind- bile devices and servers, there is a growing demand sight of x86 in its design. For example, nested page for providing the benefits of virtualization for ARM- tables, not part of the original x86 virtualization hard- based devices. We present our experiences building the ware, are standard in ARM. However, there are impor- Linux ARM hypervisor, KVM/ARM, the first full sys- tant differences between ARM and x86 virtualization ex- tem ARM virtualization solution that can run unmodified tensions such that x86 hypervisor designs may not be guest operating systems on ARM multicore hardware. directly amenable to ARM. These differences may also KVM/ARM introduces split-mode virtualization, allow- impact hypervisor performance, especially for multicore ing a hypervisor to split its execution across CPU modes systems, but have not been evaluated with real hardware. to take advantage of CPU mode-specific features. This allows KVM/ARM to leverage Linux kernel services and We describe our experiences building KVM/ARM, functionality to simplify hypervisor development and the ARM hypervisor in the mainline Linux kernel. maintainability while utilizing recent ARM hardware KVM/ARM is the first hypervisor to leverage ARM virtualization extensions to run application workloads in hardware virtualization support to run unmodified guest guest operating systems with comparable performance operating systems (OSes) on ARM multicore hardware.
    [Show full text]
  • Koller Dan Williams IBM T
    An Ounce of Prevention is Worth a Pound of Cure: Ahead-of-time Preparation for Safe High-level Container Interfaces Ricardo Koller Dan Williams IBM T. J. Watson Research Center Abstract as the container filesystem is already layered at the same file- based granularity as images. Containers continue to gain traction in the cloud as However, relying on the host to implement the filesystem lightweight alternatives to virtual machines (VMs). This abstraction is not without cost: the filesystem abstraction is is partially due to their use of host filesystem abstractions, high-level, wide, and complex, providing an ample attack which play a role in startup times, memory utilization, crash surface to the host. Bugs in the filesystem implementation consistency, file sharing, host introspection, and image man- can be exploitedby a user to crash or otherwise get controlof agement. However, the filesystem interface is high-level and the host, resulting in (at least) a complete isolation failure in wide, presenting a large attack surface to the host. Emerg- a multi-tenant environment such as a container-based cloud. ing secure container efforts focus on lowering the level A search on the CVE database [13] for kernel filesystem vul- of abstraction of the interface to the host through deprivi- nerabilities provides a sobering reminder of the width, com- leged functionality recreation (e.g., VMs, userspace kernels). plexity, and ultimately danger of filesystem interfaces. However, the filesystem abstraction is so important that some have resorted to directly exposing it from the host instead of Efforts to improve the security of containers use low-level suffering the resulting semantic gap.
    [Show full text]
  • Verifying a High-Performance Crash-Safe File System Using a Tree Specification
    Verifying a high-performance crash-safe file system using a tree specification Haogang Chen,y Tej Chajed, Alex Konradi,z Stephanie Wang,x Atalay İleri, Adam Chlipala, M. Frans Kaashoek, Nickolai Zeldovich MIT CSAIL ABSTRACT 1 INTRODUCTION File systems achieve high I/O performance and crash safety DFSCQ is the first file system that (1) provides a precise by implementing sophisticated optimizations to increase disk fsync fdatasync specification for and , which allow appli- throughput. These optimizations include deferring writing cations to achieve high performance and crash safety, and buffered data to persistent storage, grouping many trans- (2) provides a machine-checked proof that its implementa- actions into a single I/O operation, checksumming journal tion meets this specification. DFSCQ’s specification captures entries, and bypassing the write-ahead log when writing to the behavior of sophisticated optimizations, including log- file data blocks. The widely used Linux ext4 is an example bypass writes, and DFSCQ’s proof rules out some of the of an I/O-efficient file system; the above optimizations allow common bugs in file-system implementations despite the it to batch many writes into a single I/O operation and to complex optimizations. reduce the number of disk-write barriers that flush data to The key challenge in building DFSCQ is to write a speci- disk [33, 56]. Unfortunately, these optimizations complicate fication for the file system and its internal implementation a file system’s implementation. For example, it took 6 years without exposing internal file-system details. DFSCQ in- for ext4 developers to realize that two optimizations (data troduces a metadata-prefix specification that captures the writes that bypass the journal and journal checksumming) properties of fsync and fdatasync, which roughly follows taken together can lead to disclosure of previously deleted the behavior of Linux ext4.
    [Show full text]
  • NOVA Introduction
    NOVA introduction Michal Sojka1 Czech Technical University in Prague, Email: [email protected] December 5, 2017 1Based on exercises by Benjamin Engel from TU Dresden. M. Sojka B4B35OSY NOVA intro December 5, 2017 1 / 16 NOVA microhypervisor I Research project of TU Dresden (< 2012) and Intel Labs (≥ 2012). I http://hypervisor.org/, x86, GPL. I We will use a stripped down version (2 kLoC) of the microhypervisor (kernel). M. Sojka B4B35OSY NOVA intro December 5, 2017 2 / 16 Table of content Prerequisites System booting Program binaries ELF headers Virtual memory management Additional information M. Sojka B4B35OSY NOVA intro December 5, 2017 3 / 16 Prerequisites What you need to know? I NOVA is implemented in C++ (and assembler). I Each user “program” is represented by execution context data structure (class Ec). I The first executed program is called root task (similar to init process in Unix). M. Sojka B4B35OSY NOVA intro December 5, 2017 4 / 16 Prerequisites Getting started unzip nova.zip cd nova make # Compile everything make run # Run it in Qemu emulator Understanding qemu invocation qemu-system-i386 -serial stdio -kernel kern/build/hypervisor -initrd user/hello I Serial line of the emulated machine will go to stdout I Address of user/hello binary will be passed to the kernel via Multiboot info data structure Source code layout I user/ – user space code (hello world + other simple programs) I kern/ – stripped down NOVA kernel I you will need to modify kern/src/ec.cc M. Sojka B4B35OSY NOVA intro December 5, 2017 5 / 16 0 0x2000 3G (0xC0000000) Remap 4G page Virtual memory User space Kernel space EIP EIPEIPEIPEIP Kernel code/data ELF header EIP User code/data User stack multiboot user/hello kern/build/hypervisor info System booting NOVA booting 1.
    [Show full text]
  • Linux Introduction
    LINUX INTRODUCTION Until recently, running Unix meant investing in a powerful workstation that cost megabucks. Linux changes all that, because it's a complete version of the Unix operating system (software that controls the basic functions of the personal computer) that runs on ordinary personal com- puters. The added fact that it's freely available and "open source" makes it all the more attractive. Linux is perfect for people who want to operate their own low-cost Internet servers, and it's ro- bust enough to satisfy the needs of many Internet service providers. Linux is a multiuser and multitasking environment, and it can access huge amounts of me mory (gigabytes) and huge amounts of disk storage space (terabytes). Linux offers virtually everything that Windows has been promising for years and may not deliver in a truly stable form for some time to come. Don't make the mistake of assuming that Linux is some kind of watered-down or underpowered Unix for the masses. Linux is Unix. POSIX certification (compliance with the industry stan- dards for Unix) makes it official that Linux can do everything that a Unix system is supposed to do. The only difference is that Linux works on a personal computer, whereas other versions of Unix run on larger workstations or mainframes. Linux is also being taken very seriously by the computer industry, with new Linux-compatible versions of popular software packages being announced every month. The Apache Web server software running on Linux platforms powers about half of all Web sites today. Even more tel- ling, Microsoft considers Linux a major threat to its Windows empire.
    [Show full text]
  • Openshift Virtualization Technical Presentation
    OpenShift Virtualization Technical presentation 1 V0000000 What is OpenShift Virtualization? 2 V0000000 Containers are not virtual machines ● Containers are process isolation ● Kernel namespaces provide isolation and cgroups provide resource controls App 1 App 2 App 3 ● No hypervisor needed for containers Guest Guest Guest OS OS OS ● Contain only binaries, libraries, and tools App 1 App 2 App 3 which are needed by the application Hypervisor Operating System ● Ephemeral Infrastructure Infrastructure Virtualization Containerization 3 V0000000 Virtual machines can be put into containers ● A KVM virtual machine is a process ● Containers encapsulate processes ● Both have the same underlying resource needs: ○ Compute ○ Network ○ (sometimes) Storage 4 V0000000 OpenShift Virtualization ● Virtual machines ○ Running in containers ○ Using the KVM hypervisor ● Scheduled, deployed, and managed by Kubernetes ● Integrated with container orchestrator resources and services ○ Traditional Pod-like SDN connectivity and/or connectivity to external VLAN and other networks via multus ○ Persistent storage paradigm (PVC, PV, StorageClass) 5 V0000000 VM containers use KVM ● OpenShift Virtualization uses KVM, the Linux kernel hypervisor ● KVM is a core component of the Red Hat Enterprise OTHER APPS libvirt Linux kernel QEMU ○ KVM has 10+ years of production use: Red Hat Virtualization, Red Hat OpenStack Platform, and RHCOS KVM RHEL all leverage KVM, QEMU, and libvirt ● QEMU uses KVM to execute virtual machines DRIVER DRIVER DRIVER ● libvirt provides a management
    [Show full text]
  • NOVA: the Fastest File System for Nvdimms
    NOVA: The Fastest File System for NVDIMMs Steven Swanson, UC San Diego XFS F2FS NILFS EXT4 BTRFS © 2017 SNIA Persistent Memory Summit. All Rights Reserved. Disk-based file systems are inadequate for NVMM Disk-based file systems cannot 1-Sector 1-Block N-Block 1-Sector 1-Block N-Block Atomicity overwrit overwrit overwrit append append append exploit NVMM performance e e e Ext4 ✓ ✗ ✗ ✗ ✗ ✗ wb Ext4 ✓ ✓ ✗ ✓ ✗ ✓ Performance optimization Order Ext4 ✓ ✓ ✓ ✓ ✗ ✓ compromises consistency on Dataj system failure [1] Btrfs ✓ ✓ ✓ ✓ ✗ ✓ xfs ✓ ✓ ✗ ✓ ✗ ✓ Reiserfs ✓ ✓ ✗ ✓ ✗ ✓ [1] Pillai et al, All File Systems Are Not Created Equal: On the Complexity of Crafting Crash-Consistent Applications, OSDI '14. © 2017 SNIA Persistent Memory Summit. All Rights Reserved. BPFS SCMFS PMFS Aerie EXT4-DAX XFS-DAX NOVA M1FS © 2017 SNIA Persistent Memory Summit. All Rights Reserved. Previous Prototype NVMM file systems are not strongly consistent DAX does not provide data ATomic Atomicity Metadata Data Snapshot atomicity guarantee Mmap [1] So programming is more BPFS ✓ ✓ [2] ✗ ✗ difficult PMFS ✓ ✗ ✗ ✗ Ext4-DAX ✓ ✗ ✗ ✗ Xfs-DAX ✓ ✗ ✗ ✗ SCMFS ✗ ✗ ✗ ✗ Aerie ✓ ✗ ✗ ✗ © 2017 SNIA Persistent Memory Summit. All Rights Reserved. Ext4-DAX and xfs-DAX shortcomings No data atomicity support Single journal shared by all the transactions (JBD2- based) Poor performance Development teams are (rightly) “disk first”. © 2017 SNIA Persistent Memory Summit. All Rights Reserved. NOVA provides strong atomicity guarantees 1-Sector 1-Sector 1-Block 1-Block N-Block N-Block Atomicity Atomicity Metadata Data Mmap overwrite append overwrite append overwrite append Ext4 ✓ ✗ ✗ ✗ ✗ ✗ BPFS ✓ ✓ ✗ wb Ext4 ✓ ✓ ✗ ✓ ✗ ✓ PMFS ✓ ✗ ✗ Order Ext4 ✓ ✓ ✓ ✓ ✗ ✓ Ext4-DAX ✓ ✗ ✗ Dataj Btrfs ✓ ✓ ✓ ✓ ✗ ✓ Xfs-DAX ✓ ✗ ✗ xfs ✓ ✓ ✗ ✓ ✗ ✓ SCMFS ✗ ✗ ✗ Reiserfs ✓ ✓ ✗ ✓ ✗ ✓ Aerie ✓ ✗ ✗ © 2017 SNIA Persistent Memory Summit. All Rights Reserved.
    [Show full text]
  • Bull System Backup / Restore N O V a SC a LE
    Bull System Backup / Restore NOVASCALE User's Guide for NovaScale Universal & Intensive REFERENCE 86 A2 73EV 04 NOVASCALE Bull System Backup / Restore User's Guide for NovaScale Universal & Intensive Software November 2008 BULL CEDOC 357 AVENUE PATTON B.P.20845 49008 ANGERS CEDEX 01 FRANCE REFERENCE 86 A2 73EV 04 The following copyright notice protects this book under Copyright laws which prohibit such actions as, but not limited to, copying, distributing, modifying, and making derivative works. Copyright Bull SAS 2008 Printed in France Suggestions and criticisms concerning the form, content, and presentation of this book are invited. A form is provided at the end of this book for this purpose. To order additional copies of this book or other Bull Technical Publications, you are invited to use the Ordering Form also provided at the end of this book. Trademarks and Acknowledgements We acknowledge the right of proprietors of trademarks mentioned in this book. Intel ® and Itanium ® are registered trademarks of Intel Corporation. Windows ® and Microsoft ® software are registered trademarks of Microsoft Corporation. UNIX ® is a registered trademark in the United States of America and other countries licensed exclusively through the Open Group. Linux ® is a registered trademark of Linus Torvalds. NovaScale ® is a registered trademark of Bull The information in this document is subject to change without notice. Bull will not be liable for errors contained herein, or for incidental or consequential damages in connection with the use of this material.
    [Show full text]
  • Optimizing Openstack Nova for Scientific Workloads
    EPJ Web of Conferences 214, 07031 (2019) https://doi.org/10.1051/epjconf/201921407031 CHEP 2018 Optimizing OpenStack Nova for Scientific Workloads 1, 1 1 Belmiro Moreira ∗, Spyridon Trigazis , and Theodoros Tsioutsias 1European Organization for Nuclear Research (CERN), CH-1211 Geneva, Switzerland Abstract. The CERN OpenStack cloud provides over 300,000 CPU cores to run data processing analyses for the Large Hadron Collider (LHC) experiments. To deliver these services, with high performance and reliable service levels, while at the same time ensuring a continuous high resource utilization has been one of the major challenges for the CERN cloud engineering team. Several op- timizations like NUMA-aware scheduling and huge pages, have been deployed to improve scientific workloads performance, but the CERN Cloud team con- tinues to explore new possibilities like preemptible instances and containers on bare-metal. In this paper we will dive into the concept and implementation challenges of preemptible instances and containers on bare-metal for scientific workloads. We will also explore how they can improve scientific workloads throughput and infrastructure resource utilization. We will present the ongoing collaboration with the Square Kilometer Array (SKA) community to develop the necessary upstream enhancement to further improve OpenStack Nova to support large-scale scientific workloads. 1 Introduction During the last 5 years, CERN has been running a large cloud infrastructure that provides an IaaS service to the CERN community. The cloud infrastructure brought several advantages when compared to the previous model, the manual allocation of physical resources. It im- proved the user responsiveness with a self-service kiosk for virtual resources, enabled cloud interfaces like OpenStack APIs and EC2 API, improved efficiency over the entire lifetime of the resources [1].
    [Show full text]