Enterprises hold different kinds of data. Some of the data held by enterprises are mission-critical, but most of them have either historical or operational value. Moreover, some types of data are accessed every day and some are not accessed at all for years. So, all data cannot be treated equal for storage purposes. Different categories of data must be stored differently. Storage parameters must also vary for different types of data. Parameters such as protection levels, performance requirements and frequency of use may determine how and where particular categories of data are stored. 

Tiering is a phase II activity in implementation of Information Lifecycle Management (ILM). It is one method of establishing a hierarchy of storage management (HSM) systems based on business requirements. Inactive or expired data can be relegated to less expensive storage media in ILM or HSM, while mission-critical information must be stored on expensive primary enterprise arrays. Tiered storage archiving software automatically assigns different categories of data to different types of storage media, based on configurations and settings specified by the storage administrator in accordance with company policies.

Tiering resolves one of the biggest problems of data storage—reduction of storage costs. By addressing data growth and assigning different types of data to different tiers, tiering brings down the cost of storage. It improves performance throughput in backup and storage by moving less important data when backup windows are available and transmitting mission-critical data continuously or as per pre-defined schedules to online data backup servers. Tiering, thus frees up resources for mission-critical activities during peak work hours.

Data storage tiering requires certain mechanisms to be put in place. These may be static in that the mechanism assigns certain applications to certain tiers. It may be staged with data batches moving to the archive in stages or it may be dynamic with active data movers like HSM or ILM policy services, activating data movements to specific kinds of storage.

Adoption of tiered storage as a datacenter architecture is fueled by the explosion of low cost, high capacity Serial Advanced Technology Attachment (SATA) disk array systems. These systems are excellent storage options for secondary storage, online archives and compliance stores.

To summarize, tiering:

  • Automates data movements between tiers reducing OPEX
  • Optimizes use of existing infrastructure by aligning data with storage tiers and thereby reduces infrastructure costs
  • Improves performance by placing frequently accessed data in high performance tiers
  • Integrates large elements of data storage cost effectively and
  • Creates the necessary environment for information life cycle management