The Private Cloud Reference Architecture defines many key principles, concepts and patterns that must be considered when designing a private cloud infrastructure for an organization. Many of these key considerations are aided by including the Service Provider Foundation in the design of a private cloud. Service Provider Foundation, as the name implies, is primarily intended for cloud service provider organizations to enable them to build self-service portals and expose management interfaces that may be used by their customers. However large enterprise organizations share many of the same needs and concerns as cloud service providers. Consider that a large enterprise is comprised of several business entities or groups, that is they are actually tenants that require services of the IT organization. The internal IT organization is the provider of those services to the business groups. In the enterprise we have the consumer / service provider relationship that's similar to engaging with external service providers for IT resources. That introduces the private cloud principle of Take a Service Providers Approach to providing IT in large enterprise organizations. A cloud service is a shared service offering select well defined capabilities to self-service consumers. These services include the actual capability, the capacity to grow and ability to collapse as appropriate, perform as expected, and provide continuous availability. These key expected principles require the service provider to enable self-service to respond to demands of consumers and programmatic management interfaces to enable fabric management automation to respond to changes in the demand or health of services running on the infrastructure. Microsoft Windows Server 2012 and Microsoft System Center 2012 SP1 provide the platform to enable cloud infrastructures. Service Provider Foundation enables common management semantics across private and public cloud computing platforms.
Management Stamps, or stamp, is a new concept introduced with Service Provider Foundation. A stamp represents a unit of virtualized platform infrastructure that includes System Center Virtual Machine Manager, one or more virtual machine hosts and the virtual machines that are managed in the context of the System Center Virtual Machine Manager instance within the stamp. Each stamp also includes the configuration unique to each stamp such as service accounts and user roles. Stamps must be capable of being monitored; therefore a stamp also includes an instance of System Center Operations Manager. However an instance of System Center Operations Manager may provide monitoring for multiple stamps so there is not necessarily a 1:1 relationship between the number of stamps and instances of System Center Operations Manager. Put another way a stamp is an instance of System Center that supports a virtualized platform infrastructure up to the maximum number of virtual machine hosts and virtual machines supported by System Center. Stamps are an important concept since they allow service providers to distribute tenants and their services across multiple instances of System Center components (such as Operations Manager and Virtual Machine Manager), datacenters and geographic locations. Similarly they allow service designers to define how their service is deployed. For example assume a tenant of the service provider wishes to deploy two services. This service provider is a global organization with datacenters throughout the world. The tenant defines their first service requiring multiple instances that are geographically separated. The second service is similarly defined but carries an additional constraint that it only runs in European datacenters. Stamps allow the service provider to design this flexibility into their self-service portal and platform orchestration. Once requests are accepted and validated the service provider fabric management would issue the appropriate requests through the Service Provider Foundation API to deploy the services across stamps.
A tenant is an organization or user of the platform usually through creation of an account or subscription. The on-boarding of tenants will likely be defined by policy implemented and enforced by the Service Delivery Layer of the Reference Model. A tenant will have a tenant administrator role assigned to the tenant management artifact maintained by the platform. One or more users may be assigned the administrator role. Tenants are responsible for all resources that have been provisioned by the platform on behalf of the tenant and generally a metering or chargeback model exists to expose a cost structure assigned to each resource offered by the platform and chargeable to the tenant based upon usage.
This section provides an overview of scenarios enabled by System Center Service Provider Foundation. In most cases these scenarios become enabled through the use of Service Provider Framework features used in conjunction with base platform capability provided by Microsoft System Center Virtual Machine Manager and System Center Orchestrator. Again this is an overview; for more specific information on each of the System Center Service Provider Foundation features that enable a scenario refer to the product guidance available here and through links available from the scenario description or the Resources section of this article.
図2: サービスプロバイダーがIaaSを提供できるようになる
Staging Service Provider Foundation to review its capabilities and management interfaces or integrate into your development environment requires administrative access to one or more System Center Virtual Machine Manager instances that are actively managing at least one virtual machine host in a test or sandbox environment. You then need the physical or virtual machine resources to deploy System Center Orchestrator and Service Provider Foundation and configure with information about your Virtual Machine Manager stamps.
Using System Center 2012 components together as outlined in this section you should come away with an understanding of how Service Provider Foundation with Virtual Machine Manager enables Infrastructure as a Service capabilities useful to service provider and large IT organizations. More specifically you will:
In order to prepare your environment for this scenario, you should review guidance in the System Center 2012 Integration Guide hosted on the Microsoft TechNet Library. There you can review community information of each System Center component in its role as a programmable platform to be used for the Microsoft Private Cloud. It is intended to provide an abstraction layer that guides partners and customers on their decision process for methods to build automated solutions across System Center components and between System Center and other systems. Once you have the System Center 2012 components and other requirements met, you窶决e ready to explore Service Provider Foundation in your environment.
This section outlines the steps you should complete to accomplish reviewing Service Provider Foundation.
更なる技術シナリオについては、 System Center TechCenter内のTechnical Scenariosページ (英語)をご覧ください。