The SNIA announced that version 0.9 of the Cloud Data Management Interface (CDMI) is available for download.
This version incorporates comments received from the initial public review back in September, as well as some new features. One feature is the ability to access specific byte ranges of a data object. Another important feature is the ability to access specific metadata elements individually. These features will help reduce the bandwidth required to perform certain operations.
We have also incorporated feedback suggesting better conformance to HTTP and the principles of REST. The document is easier to read and better explains the concept behind the interface. One of those is the notion of tagging the containers and data objects with Data System Metadata. Table 14 in chapter 16 now goes into much more detail on the various metadata tags that are standardized:
By tagging a data object or container of data objects with this metadata, specific data requirements are expressed to the cloud. If the cloud provider can deploy data services to meet these requirements, they can charge more as a result, leading to cloud storage that is better than the typical “best effort” services offered today.
For each of these elements, there is a corresponding “billed” value that the cloud provider supplies indicating how well the data requirement has been met. The better the service provided, the more they can charge as a result. This metadata is preserved in the serialized data format that is also standardized by CDMI so that as you move data from one cloud to another (Public<->Public, or Private<->Public), the data requirements can be interpreted interoperably by the new cloud.
There are many other new features in this version of CDMI, including a new security writeup. Please download and review this version and provide feedback over the next month so we can have a robust version 1.0 early next year.