Shulou(Shulou.com)06/02 Report--

Introduction to OpenStack-Theory part (2): preface to node types and architecture of OpenStack (including sample dashboard interface)

​ read some blogs' descriptions of OpenStack architecture on the Internet, most of which were taken from the official website (or pictures described in pure English) or directly translated into Chinese and copied and pasted. If for beginners, this may be a book of heaven. So the author's previous article introduced the basics of OpenStack and the core component services. This paper first describes the node structure of OpenStack deployment, and then describes its overall architecture.

I. OpenStack node type

Before ​ introduces the node types of OpenStack, let's add some knowledge to OpenStack.

​ OpenStack is an open source cloud computing platform for all types of clouds with the goal of being easy to implement, massively scalable, and feature-rich.

​ OpenStack provides IaaS (Infrastructure as a Service) solutions through a set of interrelated service components. We all need to understand the various types of nodes and the corresponding installed services before we deploy the OpenStack practice in a lab or production environment.

There are four node types for ​ OpenStack:

Control node; network node: computing node; storage node

Let's pay attention to the introduction of these four node types.

1.1 Control Node of OpenStack

The ​ control node includes management support services, basic management services and extended management services.

1.1.1 Management support Services (underlying)

Compared with OpenStack, ​ belongs to the underlying software installation and is used to provide services to related components. For example, install database software, message queuing software, Memorycache and Etcd, etc.

1.1.2 basic Management Services (Core)

​ this service mainly installs some core services, such as Kystone, Glance, Nova, as well as Neutron and Horizon services. Used to provide user-based control and management.

1.1.3 extend management services (optional)

​ extension Management Service is an optional installation of some components, such as storage management, database services, etc. Includes major optional services for Cinder, Swift, Trove, Heat, and Ceilometer.

​ in addition, the control node generally needs to communicate or manage with each node through the network port.

1.2 OpenStack Network Node

Network nodes have and only have Neutron services, that is, network services.

Neutron is mainly responsible for managing the communication between private and public network segments, as well as the communication between virtual machine networks and firewalls.

Generally, more than two network ports are deployed during deployment, which are used to communicate with the control node, with the computing / storage node, and between the external virtual machine and the corresponding network.

1.3Compute nodes for OpenStack

Computing nodes mainly include computing services, network services and monitoring services. Basic services are the first two, and monitoring can expand the choice.

1.4 OpenStack storage node

​ storage includes block storage and object storage. Although both of them are storage, they are very different in nature.

​ block storage is equivalent to a virtual disk, which can be mounted on the corresponding virtual machine without being affected by the file system. It is generally used to expand the capacity of the scenario when the virtual machine space is insufficient. Object storage, in essence, is a virtual disk space that can store files, but cannot format or convert the file system. It is generally used for cloud disks / files.

II. The overall structure of OpenStack

​ on the premise that we know something about the types of nodes on which OpenStack is deployed in the experimental environment or production environment, let's briefly explain the overall architecture of OpenStack from two aspects, which are intercepted from official documents and divided mainly from conceptual and logical aspects.

​, of course, we still have to follow the relevant information of the official documents, but here the author will describe the whole content in detail. After all, this article provides a more detailed description and introduction to some people who have the foundation of network, cluster and linux operating system and want to get started with OpenStack.

2.1 Conceptual Architecture

​ is shown in the following figure, which shows the relationship between OpenStack services. I call it the overall concept map of OpenStack. It shows the typical projects of OpenStack very well and connects the various services. Although we may not be able to fully understand it from the beginning, we can have a better understanding of the concept of OpenStack as a whole through this diagram.

​, so how do we look at this conceptual architecture?

​ well, you can take a few minutes to take a look at this architecture diagram first, combined with the content of the previous article to see how many are some impression. Then I went to search for relevant introductions. At first, the author also went to the official website to read the official documents through the Google browser, and then found that it was not very easy to understand, then went to find materials, listen to lessons, and so on. Most of the blog articles found on the Internet intercepted the pictures of the official website and copied the official brief instructions. If you are too lazy to search, then listen to the author's understanding. Of course, the author's ability is limited, if there are any mistakes, I hope you can leave a message below for correction.

​ the author will mark the above image to facilitate introduction, please refer to the following figure:

​ first of all, on the whole, we should be familiar with the color-added part of the figure above. The various projects we call OpenStack also correspond to various services. In the following description, services and projects can be regarded as the same meaning.

​, let's first look at the red part in the middle. VMs represents a virtual machine instance. If you have played or experienced some of Aliyun's services, you will have a certain understanding of this. In fact, for OpenStack, virtual machine instance is to provide virtual services for customers, including CPU, memory, bandwidth, storage and other resources. These instances need to be supported by various OpenStack services, including providing corresponding resources and managing them.

​ next we look at the red box, the services here have a common feature-overall. For example: Keystone, I know that this service provides authentication, and there are many places (services) that need authentication, and so is the monitoring provided by Ceilometer services and the visual interface operation management provided by Horizon services. The reason why these three services are labeled with the same color block as the following Swift service (object storage) is because the four component services are less core than the green block. Of course, this is only the author's understanding, in fact, the above figure shows a variety of component services are common and commonly used, we still need to know more.

​ next let's take a look at the purple box. In addition to the virtual machine instances and object storage services described earlier, there are four core projects (green blocks), as well as four other projects.

​ core project description: Glance provides mirroring, Nova provides computing services (for example, data processing examples through Nova, etc.), Neutron provides network services (communication between each node, communication between internal and external networks, and communication with internal components (with corresponding API interface). Cinder provides storage, which is equivalent to a disk.

Introduction to the remaining four projects of ​:

​ Ironic project, bare metal, remember the Bare Metal we mentioned in the last article? In fact, the bare metal project is a special project of OpenStack, which is created for some special scenarios, mainly to specify that operations are performed on one or more bare machines. For example, if multiple physical machines are deployed at the same time when deploying big data cluster, Ironic can achieve rapid delivery of hardware infrastructure resources.

​ in addition, in some special application scenarios, users need to directly use physical server resources to meet specific needs of the feasibility and efficiency, so this is one of the reasons why OpenStack supports this project.

The ​ Trove project is a database as a service project provided by OpenStack. The main purpose is to enable users to make use of the functions of relational databases quickly and easily without the burden of dealing with complex management tasks. Cloud users and database administrators can configure and manage multiple database instances as needed.

The ​ Heat project, an orchestration service provided by OpenStack, is used to generate running cloud applications and describe cloud applications by running OpenStack API calls. Using templates, you can create most OpenStack resource types, such as instances, floating IP, volumes, security groups, and users. It also provides advanced features such as instance high availability, instance auto scaling, and nesting stacks. This enables OpenStack core projects to gain a larger user base.

The ​ Sahara project, a data processing service provided by OpenStack, mainly provides users with a simple way to pre-configure data processing frameworks (such as Hadoop,Spark and Storm) on OpenStack. This can be done by specifying configuration parameters such as framework version, cluster topology, node hardware details, and so on.

​, of course, here is mainly a brief introduction to the corresponding functions of some of the above projects and the relationship between the two, mainly to understand the relationship between the services, that is, to understand the translation of black English words and phrases, so as to understand the relationship between the various service items.

If ​ is not familiar with the English above or still thinks it is abstract, you can refer to the simplified architecture below, which may be convenient for you to understand.

​

​ above is the author's introduction to the official conceptual framework of OpenStack. The important component services involved will be described in more detail one by one later.

​, let's take a look at the official logical architecture diagram.

2.2 logical architecture

​ Let's talk about the logical architecture of OpenStack. The official logical architecture (common but not unique) is shown in the following figure:

​

​ from the above figure, we can find that the similarity with the conceptual architecture is that the corresponding component services are still there, but the difference is also very obvious. the logical architecture has an intuitive distinction between intranets and extranets, and each component service is separated by a dotted wireframe, each frame contains some storage of the corresponding component services, functional modules and API calling interfaces, etc., while the various components are connected by implementation. Represents the logical relationship between the various component services.

​, however, if the detailed description of the logical architecture is really boring, it may still feel complex and abstract, and the author said earlier that this is also a common architecture given by the government, not the only one. The following is a simplified version of the logical architecture diagram and introduction.

​ users or managers can use or manage OpenStack cloud platform systems through the network. There are three main methods shown in the figure above: command line interfaces (calls to relevant component interfaces), cloud management tools such as Rightscale and Enstratius, and graphical interface tools such as Dashboard, Cyberduck, and so on. The author has used the Web graphical interface tool provided by the component Dashboard,Horizon project provided by OpenStack to create and manage the cloud platform. It is easier to understand if you have used Aliyun related services. At the end of this paper, the author will give the Web interface diagram of the local login of the multi-node OpenStack cloud platform previously built and deployed on the laptop virtual machine.

In ​, within OpenStack, each component service has its own logical architecture, which is generally composed of functional modules, databases, or API interfaces that correspond to backends and interact with other components. These interfaces communicate through message queues, and some technical terms and nouns will be sorted out later.

​ therefore, understanding the logical architecture of OpenStack can be understood from the whole to the part and then to the individual.

III. OpenStack dashboard diagram

​ below shows how to log in to the OpenStack management interface, that is, the dashboard UI interface provided by the Horizon project.

1. Enter the ip address of the local control node in the web page and enter the OpenStack dashboard login interface.

2. Query login user name and password (log in as administrator here)

3. Log in with the picture above to enter the dashboard web management interface.

IV. Summary

​ this paper mainly introduces the architecture of OpenStack, briefly describes it from two aspects of concept and logic, and finally gives an example diagram of the Web interface provided by the Horizon project when the author deploys OpenStack.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.