Copan spotlights persistent data

By Dave Simpson

—Copan Systems, which made its initial mark by introducing the industry to the MAID acronym, this week launched another acronym—PDSA, or Persistent Data Storage Architecture—as well as an "ecosystem" of partners to shine more light on the storage requirements for persistent data.

In addition to Copan, the following vendors have lined up behind the PDSA initiative and will supply software/hardware for integrated systems: FalconStor Software, Index Engines, Oracle, Revivio, Seagate, Seven Ten Storage Software, Sun, Symantec, WysDM Software, and Xiotech. Copan has existing reseller/OEM relationships with some of those vendors. For example, Copan resells FalconStor's software, and Xiotech OEMs Copan's MAID (Massive Array of Idle Disks) systems.

Although some of the partners' products have already been integrated into the PDSA scheme, others will be integrated over the next six months.

The components of PDSA include the following:

  • Persistent data-aware applications (e.g., data protection, archive, content management);
  • A choice of interfaces, including file protocols (e.g., CIFS, NFS), native disk block protocols, and virtual tape; and
  • An API for configuration, management, and monitoring.

In addition, PDSA storage services will include compression, replication, indexing and classification, provisioning, encryption, and data migration.

Roger Archibald, senior vice president of marketing and business development at Copan, says that the PDSA initiative was driven in part by the rapidly escalating amount of persistent data at many organizations. For example, some research firms estimate that as much as 80% of corporate data can be characterized as persistent while only 20% is characterized as transactional data.

Persistent data is data that has aged to a point that it is unlikely to change or be modified, yet it needs to be retained for long periods of time. Transactional data, on the other hand, is dynamic, continuously updated, requires high-speed random read/write access, and is typically based on structured data. Databases are the primary example of transactional data. Archibald argues that most existing disk systems are tuned for transactional processing and are not optimized for storing persistent data, including content-addressed storage (CAS) systems from the large vendors, which he says are typically CAS front-ends on systems that were designed for transactional processing, albeit with Serial ATA (SATA) drives instead of Fibre Channel drives.

In contrast, Copan highlights the potential advantages of its MAID architecture in the context of storing persistent data, which are mostly in the areas of lower footprint, power, and cooling requirements, as well as reduced operational costs. (In a MAID architecture, the majority of disk drives are powered off at any time, and drives are powered on as needed when an application requests data.)

This article was originally published on September 20, 2006