Aodh
ceilometer a lot of optimizations, which separate the three characteristics of change is relatively large sample data, sampling data storage, alarm functions
these three characteristics into three independent projects, sampling data is the responsibility of the ceilometer, sampling data storage responsibility of the gnocchi, alarm function by the aodh responsible for
billing portfolio ceilometer, steady hand gnocchi, aodh, panko project Barbican
is to provide key management functionality for any environment that includes cloud services, including.
Barbican is capable of storing the "secret" includes:
key. A symmetric key algorithm, the private key asymmetric algorithms (including public key), the digital signature keys and the like. Its underlying support different hardware and software (HSM) designed by the plugin.
certificate. A certificate issued by the plugin way, verification and other functions. This feature will be deprecated in Pike version.
Binary data.
Barbican achieve plug-in extensions through the stevedore. According to the functional division, Barbican, there are three:
Crypto. Used to encrypt information to be stored. Simple_crypto and back-end support HSM. simple_crypto master key is stored in the barbican configuration file, so great a security risk.
secretstore. It may generate and save key, depending on the plug support, can generate different types of keys, including symmetric / asymmetric. Currently supported plug-ins have dogtag and kmip.
certificate. Back-end support snakeoil, symantec and dogtag. Bifrost
Bifrost 是一套 Ansible 工具，它可以在一套已知的硬件上通过 Ironic 上自动部署一个基本的镜像。它为一次性操作系统提供了模块化的实用程序，并且尽可能地减少了操作。整个过程被分为三个步骤：
安装：通过下载和建立机器镜像来准备本地环境，然后安装和配置必要的服务。
注册：输入定制化的硬件清单文件，用 Ironic 注册硬件列表，配置之前下载好的镜像
部署：指导 Ironic 将操作系统部署到每个机器上
Blazar
Blazar 是 OpenStack 的资源预留项目，通过 Blazar 用户可以让 OpenStack 在 “租约（leased）” 期内预留出特定的资源以供使用。
预留的资源类型：
虚拟预留资源：Nova Instances、Cinder Volumes、Neutron Networks
物理预留资源：Compute Host（full hosts with specific characteristics of RAM, CPU, etc）
应用场景:
为尖峰负载准备资源
将租约作为计量结算单元
优化能源消耗
申请专用资源
Ceilometer
Celiometer是OpenStack的计量与监控组件，官方的正式名称为OpenStack Telemetry，用来获取和保存计量与监控的各种测量值，并根据测量值进行报警。同时这些保存下来的测量值也可以被第三方系统获取，用来做更进一步的分析、处理或展示。
计量与监控是公有云运营的一个重要环节，计量是为了获取系统中用户对各种资源的使用情况，监控是为了确保资源处于健康的状态。
Cinder
块存储服务，提供相应的块存储，简单来说，就是虚拟出一块磁盘，可以挂载到相应的虚拟机之上，不受文件系统等因素影响，对虚拟机来说，这个操作就像是新加了一块硬盘，可以完成对磁盘的任何操作，包括挂载，卸载，格式化，转换文件系统等等操作，大多应用于虚拟机空间不足的情况下的空间扩容等等
CloudKitty
可以完成虚拟机实例(compute)、云硬盘(volume)、镜像(image)、网络进出流量(network.bw.in, network.bw.out)、浮动IP(network.floating)的计费
包括计费服务的对象获取（Tenant Fetcher）、计费数据源的收集（Collector）、计费引擎（Rating）的实现，计费费用数据的存储（Storage）
Congress
Congress是一个基于异构云环境的策略声明、监控、实施、审计的框架（policy-as-a-service）。Congress从云中不同的服务获取数据，输入到con
gress的策略引擎，从而验证云中的各服务状态是否按照设置的策略运行
Congress的功能
监控云中的策略冲突（monitoring the cloud for policy violations）
预防策略冲突（preventing violations before they occur）
矫正策略冲突（correcting violations after they occur）
Cyborg
随着机器学习和机器视觉的快速发展，用户对GPU的需求也日益剧增。截止目前，大多数用户仍会选择带有GPU的裸机服务器。然而，这同时意味着用户需要承担由配置此类设备所带来的管理性成本。如今，用户将能够使用vGPU驱动的虚拟机，并利用这部分资源运行人工智能相关的Workload。
随着OpenStack社区对AI和边缘计算的布局，而加速计算在边缘比在数据中心更为普遍，所以这又会加强OpenStack的地位，因此OpenStack在第17个版本迎来了Cyborg项目。
Cyborg项目起源于NFV acceleration management以及ETSI NFV-IFA 004 document，和OPNFV DPACC项目。Cyborg（以前称为Nomad）是用于管理硬件和软件加速资源（如 GPU、FPGA、CryptoCards和DPDK / SPDK）的框架，在Queens发布中首次亮相。特别是对于有 NFV workload的运营商，计算加速已经成为云虚拟机的必备功能。通过Cyborg，运维者可以列出、识别和发现加速器，连接和分离加速器实例，安装和卸载驱动。它也可以单独使用或与Nova或Ironic结合使用。Cyborg可以通过Nova计算控制器或Ironic裸机控制器来配置和取消配置这些设备。
在加速器方面，Nova计算控制器现在可以将Workload部署到Nvidia和Intel的虚拟化GPU（AMD GPU正在开发）。加速器可用于图形处理的场景（如虚拟桌面和工作站），还可以应用于集群上的通过虚拟化GPU以运行HPC或AI Workload的场景。
Cyborg API---应该支持有关加速器的基本操作，API支持以下接口：
attach：连接现有的物理加速器或创建新的虚拟加速器，然后分配给虚拟机
detach：分离现有物理加速器或释放虚拟机的虚拟加速器
list：列出所有附加的加速器
update：修改加速器（状态或设备本身）
admin：CRUD操作无关的某些配置
Cyborg Agent---Cyborg agent将存在于计算主机以及可能使用加速器的其他主机上，agent具体的作用：
检查硬件以找到加速器
管理安装驱动程序，依赖关系和卸载驱动
将实例连接到加速器
向Cyborg服务器报告有关可用加速器，状态和利用率的数据
硬件发现：每隔数秒就会扫描实例的加速器和现有加速器的使用级别，并将这些信息通过心跳消息报告给Cyborg服务器，以帮助管理调度加速器
硬件管理：Ansible将用于管理每个加速器的配置文件和加速器的Driver。install和uninstall特定的ansible playbook适配Cyborg所支持的硬件。在管理的硬件上进行的配置更改将通过运行不同配置的playbook作为底层实现。
实例连接：一旦产生一个实例需要连接到主机上的特定加速器，Cyborg服务器将向Cyborg agent发送消息。由于不同加速器之间的连接方法不同，因此agent需要不同的driver提供连接功能。
Cyborg-Conductor---Cyborg-db的数据库查询更新操作都需要通过向Cyborg-conductor服务发送RPC请求来实现，conductor负责数据库的访问权限控制，避免直接访问数据库。
openstack-Cyborg-generic-driver功能：
识别和发现附加的加速器后端列出在后端运行的服务
将加速器附加到通用后端
从通用后端分离加速器。
列出附加到通用后端的加速器。
修改附加到通用后端的加速器。
Quata---cyborg resource quota，Cyborg的配额管理用于在构建虚拟机时管理用户或项目对加速器的访问。目前，项目或用户可能拥有无限数量的加速资源，应该有一个限制，限制是可配置的。
Designate
提供了DNSaaS（DNS即服务）的功能，其目标就是要赋予OpenStack提供这种云域名系统的能力，云服务商可以使用Designate就能够很容易建造一个云域名管理系统来托管租户的公有域名。
，引入了pool manager机制，对DNS服务器进行池化管理。同时通过MiniDNS进行DNS区域和资源记录同步
Freezer
Freezer是一套开源的备份软件，它能帮助你自动的进行数据备份和还原动作。
目前Freezer已正式引入OpenStack，从事数据备份，是OpenStack社区中一个官方项目，旨在为OpenStack提供数据备份环境的解决方案
参考网址: https://blog.csdn.net/zhongbeida_xue/article/details/78458849
Glance
镜像服务,用来注册、登陆和检索虚拟机镜像
Heat
OpenStack is provided automatic scheduling function components, based on the template descriptive, orchestrates complex cloud applications.
Heat provides a method to easily create and manage a group of related resources to OpenStack developers and system administrators, and an orderly and predictable manner and update their resource allocation.
Heat the user can use the sample templates or create your own templates to describe any associated dependencies needed to run OpenStack resources and applications or run-time parameters.
When the setting is completed, the authorization may be controlled by pressing, predictable manner OpenStack modify and update resources.
Users can set up and update a template and its associated set of resources by OpenStack management console, Heat command-line tool or API. Horizon
provides support page Ironic
OpenStack bare metal services Ironic supports the addition of a physical machine, delete, installation and deployment of power management, so that the hardware infrastructure resources of minutes fast delivery possible. In addition, you can also customize the RAID. Even more exciting is that you can also support multi-tenant, let virtual machines communicate in the bare metal and VPC. Karbor
Karbor (formerly known as Smaug) to provide application data protection services in a OpenStack project, so that each manufacturer's data protection software through a standard interface to access OpenStack, as OpenStack provide enhanced backup, replication, migration and other data protection as a service (Data Protection as a service) capability, Karbor difficult to address virtual machine backup, the backup status of non-standard interface Keystone
is responsible for managing OpenStack framework authentication token service rules and service functions of the module. Kuryr
Kuryr projects under the OpenStackbigtent, the purpose is to network with openstackNeutron vessel docking. Kuryr gives the first impression is: this is a project under the framework of Neutron, can be controlled SDN project container network to the interface by Neutron unified North. But in fact, Kuryr Neutron is a southbound interface to docking with the container network Magnum
Mangum It is now more popular OpenStack inside a Docker and integration of new projects. Magnum is a new special project for a new Container, the container used to provide services to users
Manila
File Share Service, namely file-sharing services. OpenStack is one of the sub-project under the big tent mode, used to provide file sharing on the cloud, support CIFS and NFS protocol agreement. Mistral
WorkFlow AS A Service. Typical use cases include the Task Scheduler service cloud platform (Cloud Cron), task scheduling (Task Scheduling), long operating time of complex business process services Monasca
Monasca monitoring for the OpenStack provide a service. It is scalable, fault tolerance, support for multi-tenant integration Keystone You can follow any sampling frequency you like the index implantation. You can be embedded into existing OpenStack release, it will collect logs and metrics, applies not only to control your plane, also applies to the tenant workloads. Murano
Murano is OpenStack's Application Catalog service, respected concept AaaS (Anything-as-a-, to achieve rapid application deployment and application life cycle management functions through a unified framework and API, reduce application to the underlying platform ( and virtualization layer dependent OpenStack layer). Neutron Nova Octavia
openstack octavia is openstack lbaas a background program support, providing load-balancing virtual machine traffic. Is essentially similar to the trove, the api calls nove and neutron generating a haproxy and keepalived installed virtual machine software and connect to the target network. There are four components octavia housekeeping, worker, api, health- manager, octavia agent. api effect not explained in detail. worker: a main role and a communication component nova, neutron, etc., and for scheduling the virtual machine to the virtual machine operating instruction issued octavia agent. housekeeping: View octavia / controller / housekeeping / house_keeping.py that its features three points: SpareAmphora, DatabaseCleanup, CertRotation. Followed by cleaning up the virtual machine pool, cleaning expired database, updated certificate. health-manager: Check the virtual machine state, and a communication octavia agent virtual machine to update the status of the various components. octavia agent located inside the virtual machine: the next instruction operation is accepted haproxy underlying software, and the health-manager communication is to report each case Panko
Panko storage service providing event Rally
Tempest frame Openstack used as functional testing, and Rally it is usually used as the main performance testing and analysis tool Sahara
big data services Searchlight
purpose Searchlight is to optimize capacity and performance of the search, providing user queries for different OpenStack cloud services Senlin
Provides cluster services (Cluster as a Service) for the OpenStack cloud
the collection of objects of the same OpenStack called Cluster (Cluster). Cluster consists of nodes (node) components. Each node in the same cluster are created out of the same Profile. Note that here the nodes and cluster are abstract objects (Object), specifically what can be customized and extended with plug-ins (Plugins). For example, earlier versions of Senlin, node supports only Nova Server (ie VM Nova created) and Heat Stack (ie, contains a series of storage and network computing resources deployed by the Heat). Added support for the new year docker, that node can be a Container. Follow-up may also increase the management of the bare-metal-based Ironic to achieve.
In addition Profile scalable, the cluster nodes and management strategy (Policy) also support the expansion (based on the same mechanism Plugins). Currently Senlin project has been supported by strategies include: Deployment (including region-level and zone level), deletion, expansion, load balancing, health management. Specifically as shown below: Solum
Murano is a App Store (application store) service, and may be Solum application development App Murano posted to the Swift
object storage service, to provide the corresponding object store, simply put, is a virtualized disk space, this space which can be stored in the file, the file can be stored only can not be formatted, convert the file system, mostly applied cloud disk / file
storage node contains a minimum of two network interface Tacker
the OpenStack VNF and arrangement of the NFV management
Tacker mainly in the following modules and functions:
VNF Manger
为用户提供VNF Catalog。VNF Catalog存储着由最终用户上传的VNF descriptor（描述符？）。
对NFV实例进行基本的生命周期管理
有利于VNF的初始配置
NFV编排
提供基于端到端部署的 VNFs 平台模板
提供了横跨 VIMs 编排 VNFs灵活性
Tacker通常需要TOSCA模板作为输入 VNF管理器以编排 VNFs。TOSCA模板构成 VNF descriptor和网络服务descriptor，目前 Tacker 仅支持VNF descriptor。
Tacker的Health Monitoring （健康监测）
当前版本的Tacker的监测是去Ping 运行NFV的虚机IP，Tacker的开发者将为下一代的监测框架努力，包括：
即插即用的驱动程序来监测虚机
提高对更多事件和行为的监测驱动可用性
当Tracker无法到达虚机时，将自动重启
Tempest
Tempest 是一个旨在为云计算平台 OpenStack提供集成测试
Trove
Openstack Trove是openstack为用户提供的数据库即服务(DBaaS)。所谓DBaaS，即trove既具有数据库管理的功能，又具有云计算的优势。使用trove，用户可以：
"按需"获得数据库服务器
配置所获得的数据库服务器或者数据库服务器集群
对数据库服务器或者数据库服务器集群进行自动化管理
根据数据库的负载让数据库服务器集群动态伸缩
Vitrage
Vitrage是openstack里面提供根因分析(RCA)服务的组件。
用来组织、分析和扩展openstack的告警和事件，对问题产生的根本原因进行推导，为系统产生推导后的告警或者设置推导后的状态。
功能简介：
1.物理-虚拟实体映射
2.推导告警和状态（例如：基于系统分析后产生告警或者修改状态，而不是直接监控）
3.对告警、事件进行根因分析
4.Horizon显示
Watcher
Watcher是OpenStack中提资源优化服务组件
Watcher提供一个完整的优化循环链：从度量接收器，到优化处理器和操作计划应用程序。Watcher的目标在于提供一个强大的框架，可以实现广泛的云优化目标，包括减少数据中心运营成本，通过智能虚拟机迁移提高系统性能，提高能源效率等。此外，Watcher可供用户定制丰富的资源优化目标与策略算法
Zun
Zun是Openstack中提供容器管理服务的组件，Zun的目标是提供统一的Openstack API用于启动和管理容器，支持多种容器技术。Zun原来称为Higgins，后改名为Zun。Zun计划支持多种容器技术，Docker，Rkt，clear container等，目前只支持Docker。OpenStack Queens版本发布，由于容器社区的火热，一项值得关注的补充则为“Zun”，它在OpenStack项目中负责提供容器服务，旨在通过与Neutron、Cinder、Keystone以及其它核心OpenStack服务相集成以实现容器的快速普及。通过这种方式，OpenStack的原有网络、存储以及身份验证工具将全部适用于容器体系，从而确保容器能够满足安全与合规性要求。