In Today's post, we will be reviewing OpenStack. This is an opensource software stack for Cloud that works on Linux and Debian. It has the following components:
Nova for the computing fabric over the commodity hardware. It's a IaaS offering which can work with virtual servers, servers, storage, load balancers, and other infrastructure. It facilitates concurrent programming, database access and messaging
Swift for object storage. It promotes redundancy by storing objects and files on different computers.
This is a replacement for CloudFiles.
Cinder is one layer below. It is a block storage system that can work with different storage platform from vendors.
Neutron is a networking stack that features network configurations by user and app groups.
Horizon provides a GUI dashboard to automate cloud based resources.
Keystone provides an identity service for authentication and integrates with LDAP. AD integration is provided by some storage platform vendors.
Ceilometer provides all the instrumentation or telemetry for billing.
Heat provides an app management framework using REST and Query API.
Trove works with traditional databases.
Sahara provides an elastic map reduce framework that can work with Hadoop.
The OpenStack shared services consist of the Compute Networking and Storage stacks and these can operate on the Hypervisor and Standardized Hardware. One of the benefits of the OpenStack is that you need not know which stack is comprised of what and what vendor is providing it. The APIs abstract those away and provide a reliable set to work with. They consist of the following:
Block Storage Service API In the OpenStack Swift architecture, the storage consists of three components the Object Storage Service, the container storage service and the account storage service. The account layer process handles requests regarding metadata and the individual account The container server processes handles requests regarding container metadata or object list. The object server process is responsible for the actual storage of objects on the drives of it's node. A proxy is used to compute a hash of the storage location. An object ring is a modified consistent hashing ring that enables an account/object/container path to be mapped to partitions.
Compute API A compute worker manages computing instances on host machines.and includes commands to run, terminate, reboot instances, attach/detach volumes etc.
Identity Service API enables provisioning certificates for PKI, integrating with LDAP, token binding, user CRUD with logging and monitoring etc.
Image Service API Images are created small and are untouched. Image and runtime state are used to create instances. cinder-volume is mapped separately.
Networking API: networking services can run on different hosts and includes a cloud controller host, a network gateway host, and a number of hypervisors for hosting virtual machines.
Object Storage API such as https://swift.example.com/v1/account/container/object
Nova for the computing fabric over the commodity hardware. It's a IaaS offering which can work with virtual servers, servers, storage, load balancers, and other infrastructure. It facilitates concurrent programming, database access and messaging
Swift for object storage. It promotes redundancy by storing objects and files on different computers.
This is a replacement for CloudFiles.
Cinder is one layer below. It is a block storage system that can work with different storage platform from vendors.
Neutron is a networking stack that features network configurations by user and app groups.
Horizon provides a GUI dashboard to automate cloud based resources.
Keystone provides an identity service for authentication and integrates with LDAP. AD integration is provided by some storage platform vendors.
Ceilometer provides all the instrumentation or telemetry for billing.
Heat provides an app management framework using REST and Query API.
Trove works with traditional databases.
Sahara provides an elastic map reduce framework that can work with Hadoop.
The OpenStack shared services consist of the Compute Networking and Storage stacks and these can operate on the Hypervisor and Standardized Hardware. One of the benefits of the OpenStack is that you need not know which stack is comprised of what and what vendor is providing it. The APIs abstract those away and provide a reliable set to work with. They consist of the following:
Block Storage Service API In the OpenStack Swift architecture, the storage consists of three components the Object Storage Service, the container storage service and the account storage service. The account layer process handles requests regarding metadata and the individual account The container server processes handles requests regarding container metadata or object list. The object server process is responsible for the actual storage of objects on the drives of it's node. A proxy is used to compute a hash of the storage location. An object ring is a modified consistent hashing ring that enables an account/object/container path to be mapped to partitions.
Compute API A compute worker manages computing instances on host machines.and includes commands to run, terminate, reboot instances, attach/detach volumes etc.
Identity Service API enables provisioning certificates for PKI, integrating with LDAP, token binding, user CRUD with logging and monitoring etc.
Image Service API Images are created small and are untouched. Image and runtime state are used to create instances. cinder-volume is mapped separately.
Networking API: networking services can run on different hosts and includes a cloud controller host, a network gateway host, and a number of hypervisors for hosting virtual machines.
Object Storage API such as https://swift.example.com/v1/account/container/object
No comments:
Post a Comment