Openstack Chapter 01 -- Architecture

准备花些时间总结，学习openstack相关知识点

有兴趣的同学可以在下面评论，我会持续改进

1. openstack架构

如下图所示，openstack主要由下面三个重要的节点构成，计算，网络，控制。这和我们计算机的架构师基本相同的。

 

• The Cloud Controller. Provides all functionality of the cloud except actually hosting virtual machines or providing network services. See the "Compute Node" and "Network Controller" for details about those roles. This server will host the OpenStack Image Service, the OpenStack Block Storage Service, the OpenStack Identity Service, and the OpenStack Dashboard. It will also run portions of the OpenStack Compute service such as the API server, the scheduler, conductor, console authenticator, and VNC service. Finally, it hosts the API endpoint for the OpenStack Network service.

•  

  Dashboard and storage 都会出现在这里

• The Network Controller. Provides the bulk of the OpenStack Network services such as DHCP, layer 2 switching, layer 3 routing, floating IPs (which this guide does not configure), and metadata connectivity.

•  

  软件定义网络，SDN的概念将会在这里出现

• Compute Node. Runs the OpenStack Compute service as well as the OpenStack Network service agent (in this case, the Open vSwitch plugin agent). This server also manages an OpenStack-compatible hypervisor such as KVM or Xen. This server will host the actual virtual machines (instances).

•  

  最重要的计算节点，计算虚拟化已经是非常成熟的概念了，在KVM, VMWARE, Citrix, 这些厂商都非常成熟了，但是在openstack里面用的是hypervisor.后面继续讲解该计算虚拟化原理。

OpenStack provides great flexibility with regard to how its individual services can be hosted. For example, the services that run on the Network Controller can easily be installed on the Cloud Controller. As another example, the OpenStack Image service can be installed on its own server (or many servers to provide a more highly available service).

openstack提供了非常强大的兼容性。

(不知道有没有人玩过华为之前研发过的USM服务器，在ATAE项目组，USM关键完成了各种平台的部署，但是都是基于物理硬件的部署。而OPENSTACK是自动部署，并且可以自定义计算资源，存储资源，网络资源，等，而且具有非常强大的弹性。这就是它的强大之处。题外话)

 

Openstack 有四种网络平面

1. 管理平面

2. 数据平面

3. VM通信平面

4. API

With regard to cloud networking, a standard OpenStack Network setup can have up to four distinct physical data center networks. Note that these networks can be combined and re-used. For example, the Management, Data, and API networks are commonly the same network. For simplicity, this will be done in this guide.

• Management network. Used for internal communication between OpenStack components. The IP addresses on this network should be reachable only within the data center. 

• Data network. Used for VM data communication within the cloud deployment. The IP addressing requirements of this network depend on the OpenStack Networking plugin in use. 

• External network. Provides VMs with Internet access in some deployment scenarios. The IP addresses on this network should be reachable by anyone on the Internet. 

• API network. Exposes all OpenStack APIs, including the OpenStack Networking API, to tenants. The IP addresses on this network should be reachable by anyone on the Internet. This may be the same network as the external network, as it is possible to create a quantum subnet for the external network that uses IP allocation ranges to use only less than the full range of IP addresses in an IP block.

openstack中的数据备份API，以及数据重删功能  后续需要仔细研究一下。

在Paypal的管理团队中，能够把管理团队做的非常的细分，说明opentack在架构上设计松散。

几个关键的名词，需要在这里列一下，也可以通过这些关键词，openstack的整体框架：

OpenStack Compute(Nova)  --计算
OpenStack Object Storage(Swift) --存储
Image Service (Glance) --镜像
Identity (Keystone) --身份认证
Dashboard (Horizon) --管理平台
Network Connectivity (Quantum) --网络
Community Projects
IP Management: Melange
Load Balancer: Altas-LB -- 负载均衡
Database as a Service: RedDwarf
Message Queue: Burrow