Not Speculation

The scheme described above has historical precedents and mirrors the thinking in many development laboratories today. For example, EMC Corporation, in 2002, announced its Centera storage product that adds a "content address" data passing across its storage array controller. The purpose is to enable tracking of data through its migration between and among storage platforms over time ”and also to ensure that data has nonrepudiability (that is, that it can be demonstrated that the data has not changed bits over time or during transport).

The company's stated objective ^[10] was to create a storage solution for handling fixed content or reference data, particularly that data that must be retained online despite its low frequency of access, and with its integrity guaranteed , for regulatory reasons. University of California Berkley researchers determined in a study in 2000 (sponsored by EMC) that reference data was the fastest growing data segment in business organizations. While neither the vendor nor the university researchers had the prescience of forethought to anticipate corporate financial scandals a year or two later that brought forth a flood of new federal regulations pertaining to secure, long- term data retention, EMC has certainly profited with its Centera offering.

Centera technology is an interesting implementation of a kind of data naming scheme, but also an example of what might happen if naming schemes are not based upon open standards. To use content addressing from EMC, your storage platform must be equipped with Centera controllers, which as of this writing are only installed on EMC's own Centera storage arrays.

Data naming may not be the first objective of Microsoft Corporation, but the company has been making substantial effort to develop its brand of extended mark-up language (XML) as a data wrapper. XML is primarily viewed as a mechanism for facilitating cross-platform data exchange, replacing hard-coded electronic data interface (EDI) schemes for system-to-system data exchange. Some experimentation has been undertaken regarding the use of XML to describe objects such as files and other datasets that might be applicable to data naming going forward.

This is especially the case given Microsoft's stated direction in its next generation operating system, code named "Longhorn," to do away with file systems altogether, in favor of a SQL database and binary large objects (BLOBs). ^[11] Standardizing on a database for a file system is not a unique idea. It was proposed by Oracle Development Corporation in the late 1990s, and IBM was also talking for a while about using DB2 Open Edition and file structures as an effective file system replacement.

Going to such a database structure and replacing current file systems has the potential advantage of making implementation of a data naming scheme at the point of file creation much easier to accomplish. It also facilitates, as mentioned in a previous chapter, the reunification of storage platforms as block storage repositories, contributing to storage commoditization .