A, Keystone
1, Keystone function
(1) manage users and their rights
(2) Maintenance openstack each service Endpoint
(3) Certification and authentication
2, associated with the keystone noun: the User, Credentials, authentication, token, Project, Service, Endpoint, Role
(1) the User: the user or other service
(2) Credentials (proof): user information to prove their identity, usually a username / password or Token
(3) authentication (authentication): is the keystone to verify user identity of the process, user provides a username / password, keystone issued Token process
(4) Token: by the numbers and letters, after the user is successfully authenticated , assigned to the keystone user, the default period 24 hours
(5) Project: OpenStack for resources (computing, storage, network) and a packet separator
(6) service: computing, storage, image, and other network services, each services will offer a number of endpoint, user access resources and perform actions by endpoint
(7) Endpoint: is the address on a network accessible, usually a URL (uniform resource locator), service by exposing their API endpoint
Two, the Glance
1, the Glance architecture and component functionality (Paint)
(1) the Glance-API: provide external REST API, in response to the query image, access and call, glance-api not really handle the request stored.
(2) Glance-registry: the image processing metadata.
(3) Store backend: store the image. It supports a variety of ways storage, including file systems, Swift, http, Amazon S3 and so on.
(4) Database: Metadata image will be saved to the database, the default is MySQL.
Three, Nova
. 1, described in "creation process from the virtual machine to see how the work subservice nova- *" (Paint)
(. 1) Nova-API: receiving and responding to customer calls
(2) RabbitMQ: an information service from the relay station , coordinate communication between the sub-service.
(3) Scheduler: The resource usage of the computing nodes, select the most appropriate computing node to run the virtual machine.
Filter scheduler schedulers involves: scheduling two-step process:
A: computing node that satisfies the condition through the filter filter
B: selecting the optimal weights calculated by the weight calculation node creates a virtual machine
(4) nova-compute: compute node run virtual machines on a responsible management node.
(5) nova-conductor: nova -compute needs to obtain and update the virtual machine information database, but does not directly access the database, the database access is achieved by nova-conductor.
(6) Database: The state data storage component.
Process:
(1) the client sends a request to nova-api: "Help me create a virtual machine."
After (2) nova-api receives and processes the request, sending a message to RabbitMQ: "Let Scheduler creates a virtual machine."
(3) Scheduler RabbitMQ received request begin scheduling algorithm, selected from a plurality of computing nodes in the computing node A.
(4) Scheduler sends a message to RabbitMQ: "Creating the virtual machines on computing node A".
(5) RabbitMQ computing node A sends a message to the nova-compute, and then starts the virtual machine on the local node Hypervisor.
(6) nova-compute If you need to query and update database information, it will send a message through RabbitMQ to nova-conductor, nova-conductor responsible for database access.
Four, Cinder
. 1, Cinder structure and function components (Paint)
(. 1) Cinder-API: API receives a request, performs call cinder-volume operations.
(2) Message queue (RabbitMQ) : through the message queue, interprocess communication and mutual cooperation.
(3) Cinder-scheduler: scheduler volume created by the scheduling algorithm selects the most appropriate storage node.
(4) Cinder-volume: volume management services, and the volume provider work together to manage the life cycle of volume. Nodes running cinder-volume service is referred to as a storage node.
(5) Volume provider: providing physical storage space for the volume (create volume).
(6) Database: cinder data need to be stored in a database, typically using MySQL. The database is installed in the control node.
Five, Neutron
1, Neutron architecture and component functionality (Paint)
(1) Neutron Server: openstack provide external network API, receives the request, processes the request and call Plugin.
(2) Plugin: neutron server receives the request, call processing request Agent.
(3) Agent: receiving a request Plugin, responsible for real variety of network functions on the network provider.
(4) Network provider: to provide network services virtual or physical network devices, such as Linux Bridge, Open vSwitch virtual switch or other support Neutron physical switch.
(5) Queue: responsible for communication between the various sub-services.
(6) Database: storing network status information, including network, subnet, port, router like.