Package org.dasein.cloud.services.dc

Concepts support the physical dependencies of the cloud infrastructure.

See:
Description

Interface Summary

DataCenterServices

Describe the physical configuration of the underlying cloud provider so applications may determine how to create redundancies and protect jurisdictional concerns.

Class Summary

DataCenter	A data center represents a group of virtualized resources that share a significant amount of common physical infrastructure.
Region	A group of data centers that share a jurisidiction and some common points of failure.

Enum Summary

Jurisdiction

Represents the legal context of a given region of data centers.

Package org.dasein.cloud.services.dc Description

Concepts support the physical dependencies of the cloud infrastructure. The core class within this package is the DataCenterServices class which serves as the interface into the underlying cloud's physical geography. Each cloud is assumed to have at least one region with at least one data center.

Within the cloud, a region is any set of one or more data centers with a common jurisidiction and some level of minimal common points of failure. A consumer should generally expect the following to be true about regions:

• If one region fails entirely, other regions should remain operating (or at least be down for unrelated reasons).
• You should check business rules relating to jurisdiction before sharing data across two regions.

Notes for Implementors

Many clouds, especially private clouds, will not have the rich geographic topology of clouds like Amazon Web Services. In those cases you will likely be implementing a single region and a single data center to support this API. Where you draw the line between a region and a data center for a cloud that has no inherent concept of region or data center is up to the implementor. For example, a VMware environment with two different VMware implementations in the same rack could represent two data centers in the same region, even though there are significant common points of failure. It may, on the other hand, be more appropriate to model the two implementations as distinct clouds.