diff options
Diffstat (limited to 'docs/release/release-notes/Auto-release-notes.rst')
| -rw-r--r-- | docs/release/release-notes/Auto-release-notes.rst | 329 |
1 files changed, 267 insertions, 62 deletions
diff --git a/docs/release/release-notes/Auto-release-notes.rst b/docs/release/release-notes/Auto-release-notes.rst index eab68cc..ed6524d 100644 --- a/docs/release/release-notes/Auto-release-notes.rst +++ b/docs/release/release-notes/Auto-release-notes.rst @@ -7,31 +7,63 @@ Auto Release Notes ================== -This document provides the release notes for Fraser release of Auto. +This document provides the release notes for the Gambia 7.0 release of Auto. Important notes for this release ================================ -Initial release (project inception: July 2017). +The initial release for Auto was in Fraser 6.0 (project inception: July 2017). Summary ======= -OPNFV is a SDNFV system integration project for open-source components, which so far have been mostly limited to the NFVI+VIM as generally described by ETSI. +Overview +^^^^^^^^ + +OPNFV is an SDNFV system integration project for open-source components, which so far have been mostly limited to +the NFVI+VIM as generally described by `ETSI <https://www.etsi.org/technologies-clusters/technologies/nfv>`_. In particular, OPNFV has yet to integrate higher-level automation features for VNFs and end-to-end Services. -Auto ("ONAP-Automated OPNFV") will focus on ONAP component integration and verification with OPNFV reference platforms/scenarios, through primarily a post-install process in order to avoid impact to OPNFV installer projects. As much as possible, this will use a generic installation/integration process (not specific to any OPNFV installer's technology). +As an OPNFV project, Auto (*ONAP-Automated OPNFV*) will focus on ONAP component integration and verification with +OPNFV reference platforms/scenarios, through primarily a post-install process, in order to avoid impact to OPNFV +installer projects (Fuel/MCP, Compass4NFV, Apex/TripleO, Daisy4NFV). As much as possible, this will use a generic +installation/integration process (not specific to any OPNFV installer's technology). + +* `ONAP <https://www.onap.org/>`_ (a Linux Foundation Project) is an open source software platform that delivers + robust capabilities for the design, creation, orchestration, monitoring, and life cycle management of + Software-Defined Networks (SDNs). The current release of ONAP is B (Beijing). + +Auto aims at validating the business value of ONAP in general, but especially within an OPNFV infrastructure +(integration of ONAP and OPNFV). Business value is measured in terms of improved service quality (performance, +reliability, ...) and OPEX reduction (VNF management simplification, power consumption reduction, ...), as +demonstrated by use cases. + +Auto also validates multi-architecture software (binary images and containers) availability of ONAP and OPNFV: +CPUs (x86, ARM) and Clouds (MultiVIM) -* `ONAP <https://www.onap.org/>`_ (a Linux Foundation Project) is an open source software platform that delivers robust capabilities for the design, creation, orchestration, monitoring, and life cycle management of Software-Defined Networks (SDNs). +In other words, Auto is a turnkey approach to automatically deploy an integrated open-source virtual network +based on OPNFV (as infrastructure) and ONAP (as end-to-end service manager), that demonstrates business value +to end-users (IT/Telco service providers, enterprises). -While all of ONAP is in scope, as it proceeds, the project will focus on specific aspects of this integration and verification in each release. Some example topics and work items include: + +While all of ONAP is in scope, as it proceeds, the Auto project will focus on specific aspects of this integration +and verification in each release. Some example topics and work items include: * How ONAP meets VNFM standards, and interacts with VNFs from different vendors -* How ONAP SDN-C uses OPNFV existing features, e.g. NetReady, in a two-layer controller architecture in which the upper layer (global controller) is replaceable, and the lower layer can use different vendor’s local controller to interact with SDN-C -* What data collection interface VNF and controllers provide to ONAP DCAE, and (through DCAE), to closed-loop control functions such as Policy Tests which verify interoperability of ONAP automation/lifecycle features with specific NFVI and VIM features, as prioritized by the project with technical community and EUAG input. Examples include: +* How ONAP SDN-C uses OPNFV existing features, e.g. NetReady, in a two-layer controller architecture in which the + upper layer (global controller) is replaceable, and the lower layer can use different vendor’s local controller to + interact with SDN-C. For interaction with multiple cloud infrastructures, the MultiVIM ONAP component will be used. +* How ONAP leverages OPNFV installers (Fuel/MCP, Compass4NFV, Apex/TripleO, Daisy4NFV) to provide a cloud + instance (starting with OpenStack) on which to install the tool ONAP +* What data collection interface VNF and controllers provide to ONAP DCAE, and (through DCAE), to closed-loop control + functions such as Policy Tests which verify interoperability of ONAP automation/lifecycle features with specific NFVI + and VIM features, as prioritized by the project with OPNFV technical community and + EUAG (`End User Advisory Group <https://www.opnfv.org/end-users/end-user-advisory-group>`_) input. + + Examples: * Abstraction of networking tech/features e.g. through NetReady/Gluon * Blueprint-based VNF deployment (HOT, TOSCA, YANG) @@ -39,7 +71,8 @@ While all of ONAP is in scope, as it proceeds, the project will focus on specifi * Policy (through DCAE) * Telemetry (through VES/DCAE) -Initial areas of focus for Auto (in orange dotted lines; this scope can be expanded for future releases). It is understood that: +Initial areas of focus for Auto (in orange dotted lines; this scope can be expanded for future releases). +It is understood that: * ONAP scope extends beyond the lines drawn below * ONAP architecture does not necessarily align with the ETSI NFV inspired diagrams this is based upon @@ -47,56 +80,179 @@ Initial areas of focus for Auto (in orange dotted lines; this scope can be expan .. image:: auto-proj-rn01.png -Testability: +The current ONAP architecture overview can be found `here <https://onap.readthedocs.io/en/latest/guides/onap-developer/architecture/onap-architecture.html>`_. + +For reference, the ONAP-Beijing architecture diagram is replicated here: + +.. image:: ONAP-toplevel-beijing.png + + +Within OPNFV, Auto leverages tools and collaborates with other projects: + +* use clouds/VIMs as installed in OPNFV infrastructure (e.g. OpenStack as installed by Fuel/MCP, Compass4NFV, etc.) +* include VNFs developed by OPNFV data plane groups (e.g., accelerated by VPP (Vector Packet Processing) with DPDK support, ...) +* validate ONAP+VNFs+VIMs on two major CPU architectures: x86 (CISC), Arm (RISC); collaborate with OPNFV/Armband +* work with other related groups in OPNFV: + + * FuncTest for software verification (CI/CD, Pass/Fail) + * Yardstick for metric management (quantitative measurements) + * VES (VNF Event Stream) and Barometer for VNF monitoring (feed to ONAP/DCAE) + * Edge Cloud as use case + +* leverage OPNFV tools and infrastructure: + + * Pharos as LaaS: transient pods (3-week bookings) and permanent Arm pod (6 servers) + * `WorksOnArm <http://worksonarm.com/cluster>`_ (`GitHub link <http://github.com/worksonarm/cluster>`_) + * possibly other labs from the community (Huawei pod-12, 6 servers, x86) + * JJB/Jenkins for CI/CD (and follow OPNFV scenario convention) + * Gerrit/Git for code and documents reviewing and archiving (similar to ONAP: Linux Foundation umbrella) + * follow OPNFV releases (Releng group) + + +Testability +^^^^^^^^^^^ -* Tests will be developed for use cases within the project scope. +* Tests (test cases) will be developed for use cases within the project scope. * In future releases, tests will be added to Functest runs for supporting scenarios. -Auto’s goals include the standup and tests for integrated ONAP-Cloud platforms (“Cloud” here being OPNFV “scenarios” or other cloud environments). Thus, the artifacts would be tools to deploy ONAP (leveraging OOM whenever possible (starting with Beijing release of ONAP), and a preference for the containerized version of ONAP), to integrate it with clouds, to onboard and deploy test VNFs, to configure policies and closed-loop controls, and to run use-case defined tests against that integrated environment. OPNFV scenarios would be a possible component in the above. +Auto’s goals include the standup and tests for integrated ONAP-Cloud platforms (“Cloud” here being OPNFV “scenarios” +or other cloud environments). Thus, the artifacts would be tools to deploy ONAP (leveraging OOM whenever possible, +starting with Beijing release of ONAP, and a preference for the containerized version of ONAP), to integrate it with +clouds, to onboard and deploy test VNFs, to configure policies and closed-loop controls, and to run use-case defined +tests against that integrated environment. OPNFV scenarios would be a possible component in the above. + +Installing Auto components and running a battery of tests will be automated, with some or all of the tests being +integrated in OPNFV CI/CD (depending on the execution length and resource consumption). + +Combining all potential parameters, a full set of Auto test case executions can result in thousands of individual results. +The analysis of these results can be performed by humans, or even by ML/AI (Machine Learning, Artificial Intelligence). +Test results will be used to fine-tune policies and closed-loop controls configured in ONAP, for increased ONAP business +value (i.e., find/determine policies and controls which yield optimized ONAP business value metrics such as OPEX). + +More precisely, the following list shows parameters that could be applied to an Auto full run of test cases: + +* Auto test cases for given use cases +* OPNFV installer {Fuel/MCP, Compass4NFV, Apex/TripleO, Daisy4NFV} +* OPNFV availability scenario {HA, noHA} +* environment where ONAP runs {bare metal servers, VMs from clouds (OpenStack, AWS, GCP, Azure, ...), containers} +* ONAP installation type {bare metal, VM, or container, ...} and options {MultiVIM single|distributed, ...} +* VNF types {vFW, vCPE, vAAA, vDHCP, vDNS, vHSS, ...} and VNF-based services {vIMS, vEPC, ...} +* cloud where VNFs run {OpenStack, AWS, GCP, Azure, ...} +* VNF host type {VM, container} +* CPU architectures {x86/AMD64, ARM/aarch64} for ONAP software and for VNF software; not really important for Auto software; +* pod size and technology (RAM, storage, CPU cores/threads, NICs) +* traffic types and amounts/volumes; traffic generators (although that should not really matter); +* ONAP configuration {especially policies and closed-loop controls; monitoring types for DCAE: VES, ...} +* versions of every component {Linux OS (Ubuntu, CentOS), OPNFV release, clouds, ONAP, VNFs, ...} + +The diagram below shows Auto parameters: + +.. image:: auto-proj-parameters.png + + +The next figure is an illustration of the Auto analysis loop (design, configuration, execution, result analysis) +based on test cases covering as many parameters as possible : + +.. image:: auto-proj-tests.png + + +Auto currently defines three use cases: Edge Cloud (UC1), Resiliency Improvements (UC2), and Enterprise vCPE (UC3). These use cases aim to show: + +* increased autonomy of Edge Cloud management (automation, catalog-based deployment). This use case relates to the + `OPNFV Edge Cloud <https://wiki.opnfv.org/display/PROJ/Edge+cloud>`_ initiative. +* increased resilience (i.e. fast VNF recovery in case of failure or problem, thanks to closed-loop control), + including end-to-end composite services of which a Cloud Manager may not be aware (VMs or containers could be + recovered by a Cloud Manager, but not necessarily an end-to-end service built on top of VMs or containers). +* enterprise-grade performance of vCPEs (certification during onboarding, then real-time performance assurance with + SLAs and HA, as well as scaling). + +The use cases define test cases, which initially will be independent, but which might eventually be integrated to `FuncTest <https://wiki.opnfv.org/display/functest/Opnfv+Functional+Testing>`_. + +Additional use cases can be added in the future, such as vIMS (example: project `Clearwater <http://www.projectclearwater.org/>`_) +or residential vHGW (virtual Home Gateways). The interest for vHGW is to reduce overall power consumption: even in idle mode, +physical HGWs in residential premises consume a lot of energy. Virtualizing that service to the Service Provider edge data center +would allow to minimize that consumption. + -Auto currently defines three use cases: Edge Cloud, Resiliency Improvements, and Enterprise vCPE. These use cases aim to show: +Lab environment +^^^^^^^^^^^^^^^ -* increased autonomy of Edge Cloud management (automation, catalog-based deployment) -* increased resilience (i.e. fast VNF recovery in case of failure or problem, thanks to closed-loop control), including end-to-end composite services of which a Cloud Manager may not be aware -* enterprise-grade performance of vCPEs (certification during onboarding, then real-time performance assurance with SLAs and HA as well as scaling). +Target architectures for all Auto use cases and test cases include x86 and Arm. Power consumption analysis will be +performed, leveraging Functest tools (based on RedFish/IPMI/ILO). -The use cases define test cases, which initially will be independent, but which might eventually be integrated to FuncTest. +Initially, an ONAP-Amsterdam instance (without DCAE) had been installed over Kubernetes on bare metal on a single-server +x86 pod at UNH IOL. -Additional use cases can be added in the future, such as vIMS (example: project Clearwater). +A transition is in progress, to leverage OPNFV LaaS (Lab-as-a-Service) pods (`Pharos <https://labs.opnfv.org/>`_). +These pods can be booked for 3 weeks only (with an extension for a maximum of 2 weeks), so they are not a permanent resource. -Target architectures include x86 and Arm. +For ONAP-Beijing, a repeatable automated installation procedure is being developed, using 3 Pharos servers (x86 for now). +Also, a more permanent ONAP installation is in progress at a Huawei lab (pod-12, consisting of 6 x86 servers, +1 as jump server, the other 5 with this example allocation: 3 for ONAP components, and 2 for an OPNFV infratructure: +Openstack installed by Compass4NFV). -An ONAP instance (without DCAE) has been installed over Kubernetes on bare metal on an x86 pod of 6 servers at UNH IOL. A transition is in progress, to leverage OPNFV LaaS (Lab-as-a-Service) pods (`Pharos <https://labs.opnfv.org/>`_). -ONAP-based onboarding and deployment of VNFs is in progress (ONAP pre-loading of VNFs must still done outside of ONAP: for VM-based VNFs, need to prepare OpenStack stacks (using Heat templates), then make an instance snapshot which serves as the binary image of the VNF). +ONAP-based onboarding and deployment of VNFs is in progress (ONAP-Amsterdam pre-loading of VNFs must still done outside +of ONAP: for VM-based VNFs, users need to prepare OpenStack stacks (using Heat templates), then make an instance snapshot +which serves as the binary image of the VNF). + +A script to prepare an OpenStack instance for ONAP (creation of a public and a private network, with a router, +pre-loading of images and flavors, creation of a security group and an ONAP user) has been developed. It leverages +OpenStack SDK. It has a delete option, so it can be invoked to delete these objects for example in a tear-down procedure. Integration with Arm servers has started (exploring binary compatibility): -* Openstack is currently installed on a 6-server pod of Arm servers +* The Auto project has a specific 6-server pod of Arm servers, which is currently loaned to ONAP integration team, + to build ONAP images * A set of 14 additional Arm servers was deployed at UNH, for increased capacity -* Arm-compatible Docker images are in the process of being developed +* ONAP Docker registry: ONAP-specific images for ARM are being built, with the purpose of populating ONAP nexus2 + (Maven2 artifacts) and nexus3 (Docker containers) repositories at Linux Foundation. Docker images are + multi-architecture, and the manifest of an image may contain 1 or more layers (for example 2 layers: x86/AMD64 + and ARM/aarch64). One of ONAP-Casablanca architectural requirements is to be CPU-architecture independent. + There are almost 150 Docker containers in a complete ONAP instance. Currently, more disk space is being added + to the ARM nodes (configuration of Nova, and/or additional actual physical storage space). + + +Test case design and implementation for the three use cases has started. + +OPNFV CI/CD integration with JJD (Jenkins Job Description) has started: see the Auto plan description +`here <https://wiki.opnfv.org/display/AUTO/CI+for+Auto>`_. The permanent resource for that is the 6-server Arm +pod, hosted at UNH. The CI directory from the Auto repository is `here <https://git.opnfv.org/auto/tree/ci>`_ -Test case implementation for the three use cases has started. Finally, the following figure illustrates Auto in terms of project activities: .. image:: auto-project-activities.png +Note: a demo was delivered at the OpenStack Summit in Vancouver on May 21st 2018, to illustrate the deployment of +a WordPress application (WordPress is a platform for websites and blogs) deployed on a multi-architecture cloud (mix +of x86 and Arm servers). +This shows how service providers and enterprises can diversify their data centers with servers of different architectures, +and select architectures best suited to each use case (mapping application components to architectures: DBs, +interactive servers, number-crunching modules, ...). +This prefigures how other examples such as ONAP, VIMs, and VNFs could also be deployed on heterogeneous multi-architecture +environments (open infrastructure), orchestrated by Kubernetes. The Auto installation scripts covering all the parameters +described above could expand on that approach. + +.. image:: auto-proj-openstacksummit1805.png + + + Release Data ============ +--------------------------------------+--------------------------------------+ -| **Project** | Fraser/auto/auto@opnfv | +| **Project** | Auto | | | | +--------------------------------------+--------------------------------------+ -| **Repo/commit-ID** | | +| **Repo/commit-ID** | auto/opnfv-7.0.0 | | | | +--------------------------------------+--------------------------------------+ -| **Release designation** | Fraser 6.0 | +| **Release designation** | Gambia 7.0 | | | | +--------------------------------------+--------------------------------------+ -| **Release date** | 2018-04-20 | +| **Release date** | 2018-11-02 | | | | +--------------------------------------+--------------------------------------+ | **Purpose of the delivery** | Official OPNFV release | @@ -122,43 +278,91 @@ Reason for version Feature additions ~~~~~~~~~~~~~~~~~ -Initial release, with use case descriptions, release plan, and in-progress test cases and ONAP installations. +Initial release 6.0: +* Fraser release plan +* use case descriptions +* test case descriptions +* in-progress test case development +* lab: OPNFV and ONAP (Amsterdam) installations + +Point release 6.1: + +* added Gambia release plan +* started integration with CI/CD (JJB) on permanent Arm pod +* Arm demo at OpenStack Summit +* initial script for configuring OpenStack instance for ONAP, using OpenStack SDK 0.13 +* initial attempts to install ONAP Beijing +* alignment with OPNFV Edge Cloud +* initial contacts with Functest + +Point release 6.2: + +* initial scripts for OPNFV CI/CD, registration of Jenkins slave on `Arm pod <https://build.opnfv.org/ci/view/auto/>`_ +* updated script for configuring OpenStack instance for ONAP, using OpenStack SDK 0.14 + +Point release 7.0: + +* progress on Docker registry of ONAP's Arm images +* progress on ONAP installation script for 3-server cluster of UNH servers +* CI scripts for OPNFV installers: Fuel/MCP (x86), Compass, Apex/TripleO (must run twice) +* initial CI script for Daisy4NFV (work in progress) +* JOID script, but supported only until R6.2, not Gambia 7.0 +* completed script for configuring OpenStack instance for ONAP, using OpenStack SDK 0.17 +* use of an additional lab resource for Auto development: 6-server x86 pod (huawei-pod12) -**JIRA TICKETS:** + + + + +**JIRA TICKETS for this release:** + + +`JIRA Auto Gambia 7.0.0 Done <https://jira.opnfv.org/issues/?filter=12403>`_ + +Manual selection of significant JIRA tickets for this version's highlights: +--------------------------------------+--------------------------------------+ | **JIRA REFERENCE** | **SLOGAN** | | | | +--------------------------------------+--------------------------------------+ -| AUTO-1, UC1 definition | Define Auto-UC-01 Service Provider's | -| | Management of Edge Cloud | +| AUTO-37 | Get DCAE running onto Pharos | +| | deployment | +--------------------------------------+--------------------------------------+ -| AUTO-2, UC2 definition | Define Auto-UC-02 Resilience | -| | Improvements through ONAP | +| AUTO-42 | Use Compass4NFV to create an | +| | OpenStack instance on a UNH pod | +--------------------------------------+--------------------------------------+ -| AUTO-7, UC3 definition | Define Auto-UC-03 Enterprise vCPE | -| | | +| AUTO-43 | String together scripts for Fuel, | +| | Tool installation, ONAP preparation | +--------------------------------------+--------------------------------------+ -| AUTO-4, UC2 test case definition | Develop test cases for Auto-UC-02 | -| | Resilience Improvements through ONAP | +| AUTO-44 | Build ONAP components for arm64 | +| | platform | +--------------------------------------+--------------------------------------+ -| AUTO-8, UC3 test case definition | Develop test cases for Auto-UC-03 | -| | Enterprise vCPE | +| AUTO-45 | CI: Jenkins definition of verify and | +| | merge jobs | +--------------------------------------+--------------------------------------+ -| (UC1 test case definition is done, | | -| but no associated JIRA ticket) | | +| AUTO-46 | Use Apex to create an OpenStack | +| | instance on a UNH pod | +--------------------------------------+--------------------------------------+ -| AUTO-5, install ONAP | Getting ONAP running onto Pharos | -| | deployment (without DCAE) | +| AUTO-47 | Install ONAP with Kubernetes on LaaS | +| | | ++--------------------------------------+--------------------------------------+ +| AUTO-48 | Create documentation for ONAP | +| | deployment with Kubernetes on LaaS | ++--------------------------------------+--------------------------------------+ +| AUTO-49 | Automate ONAP deployment with | +| | Kubernetes on LaaS | ++--------------------------------------+--------------------------------------+ +| AUTO-51 | huawei-pod12: Prepare IDF and PDF | +| | files | +--------------------------------------+--------------------------------------+ -| AUTO-31, UC1 test case progress | auto-edge-pif-001 Basic OpenStack | -| | environment check | +| AUTO-52 | Deploy a running ONAP instance on | +| | huawei-pod12 | +--------------------------------------+--------------------------------------+ -| AUTO-13, UC2 test case progress | Develop test script for vif-001: | -| | Data Management | +| AUTO-54 | Use Daisy4nfv to create an OpenStack | +| | instance on a UNH pod | +--------------------------------------+--------------------------------------+ -| AUTO-20, UC3 test case progress | Onboarding of VNFs via SDC GUI | +| | | | | | +--------------------------------------+--------------------------------------+ @@ -187,18 +391,19 @@ Deliverables Software deliverables ^^^^^^^^^^^^^^^^^^^^^ -Initial release: in-progress install scripts and test case implementations. +7.0 release: in-progress Docker ARM images, install scripts, CI scripts, and test case implementations. Documentation deliverables ^^^^^^^^^^^^^^^^^^^^^^^^^^ -Initial versions of: +Updated versions of: -* Release notes (this document) -* User guide `OPNFV User and Configuration Guide <http://docs.opnfv.org/en/latest/release/userguide.introduction.html>`_ -* Configuration Guide (same landing page as User Guide) +* Release Notes (this document) +* User Guide +* Configuration Guide +(see links in References section) @@ -208,9 +413,6 @@ Known Limitations, Issues and Workarounds System Limitations ^^^^^^^^^^^^^^^^^^ -* ONAP still to be validated for Arm servers -* DCAE still to be validated for Kubernetes - Known issues @@ -260,8 +462,8 @@ None at this point. References ========== -For more information on the OPNFV Fraser release, please see: -http://opnfv.org/fraser +For more information on the OPNFV Gambia release, please see: +http://opnfv.org/gambia Auto Wiki pages: @@ -271,9 +473,9 @@ Auto Wiki pages: OPNFV documentation on Auto: -* `Auto release notes <http://docs.opnfv.org/en/latest/release/release-notes.html>`_ -* `Auto use case user guides <http://docs.opnfv.org/en/latest/submodules/auto/docs/release/userguide/index.html#auto-userguide>`_ -* `Auto configuration guide <http://docs.opnfv.org/en/latest/submodules/auto/docs/release/configguide/index.html#auto-configguide>`_ +* `Auto release notes <auto-releasenotes>` +* `Auto use case user guides <auto-userguide>` +* `Auto configuration guide <auto-configguide>` Git&Gerrit Auto repositories: @@ -282,4 +484,7 @@ Git&Gerrit Auto repositories: * `Gerrit for Auto project <https://gerrit.opnfv.org/gerrit/#/admin/projects/auto>`_ +Demo at OpenStack summit May 2018 (Vancouver, BC, Canada): + +* YouTube video (10min 52s): `Integration testing on an OpenStack public cloud <https://youtu.be/BJ05YuusNYw>`_ |