XAM - The Next-Generation API for Cloud Storage?
Frank Bunn, SNIA Europe Director and Liaison to the SNIA XAM initiative, XAM-europe-liaison@snia.org

On 6 April 2009, the SNIA Cloud Storage Technical Work Group (TWG) was launched at SNW Orlando. The aim of this group is the definition, development and market introduction of standardised interfaces for the interaction of data and storage services in a cloud environment. Thanks to the relevant work carried out in recent years by the SNIA security, storage, and data management working groups, the Cloud Storage TWG was in the fortunate position to be starting with an advantage.

Among the major challenges in Cloud Storage deployments are data classification and data migration between heterogeneous storage and applications; thanks to its metadata management capabilities, the SNIA XAM (eXtensible Access Method) specification has the potential to become the next-generation data storage interface for Cloud Storage.

Cloud storage is making headlines for its ability to provide benefits to many applications. Online backups, data archiving and e-discovery, provisioning of block- or file system-based storage on demand, as well as web-access to stored data. Because of this diversity it is already apparent that cloud storage will need to support different interfaces related to read / write, import / export, queries, metadata, and other activities. For this approach to be successful however, users should be able to eliminate the need to manage physical storage and run cloud storage-based services without the help of a specialist. In such an environment, all storage processes must be completely transparent and the administrator should only be concerned with the management of their data from the user’s perspective; the physical storage and geographical distribution based on the associated metadata should be the sole responsibility of the cloud. For the administrator, this architecture also offers the benefit of charging users based on the actual storage capacity and tiering they utilise (pay-as-you-go).

XAM is ideally suited to meet all of the above requirements. As the storage application interface, XAM standardises metadata, allowing data to be transparently migrated between applications and storage systems. The complexity of the underlying cloud layer is hidden from the user.

Graphic 1:
XAM as an interface between application and cloud storage services

From a technical perspective, the XAM API consists of three primary and two secondary objects managed by the applications during their read and write commands. An important primary XAM object in this context is the XSET, which serves both as a container for the data itself, as well as for the associated metadata. Including additional fields for classification and retention purposes, XSET metadata goes far beyond typical system metadata; XSETs themselves are dynamically created. This model is therefore very well suited for some key cloud storage requirements, such as self-service provisioning and pay-as-you go charging procedures.

Graphic 2:
XSET Object as a container for data, location and metadata

Information management and content applications add metadata to the XSET objects in order to define certain data retention policies; based on that metadata, the cloud controls the storage management.

The future interface for cloud storage must be based on a strong metadata model; XAM is ideally positioned to be this interface and therefore the foundation for XAM as the next-generation data storage interface for cloud storage has already been laid.

Back to front page..