Mending the DBA-storage manager relationship

From their perspective, storage managers see databases as somewhat of a black hole

There has always been a bit of a love/hate relationship between database administrators and storage managers. Database applications are among the largest consumers of storage within an organization and often house its most critical data. As a result, database administrators represent a key set of "customers" to storage managers. Due to the importance of their data, administrators can be somewhat demanding when it comes to storage needs -- performance and stability are paramount. This can translate into database administrators dictating storage specifications and configuration details that one party in the relationship might view as encroaching on the storage manager's turf.

From their perspective, storage managers see databases as something of a black hole -- they appear to them as large, monolithic containers with no visibility for attributes like utilization -- and they can present challenges to routine storage management functions like backup. To further complicate the relationship, most storage managers today are under significant pressure to reduce or at least contain storage growth. As a result, they've embraced concepts such as tiered storage and information life-cycle management (ILM). However, the inscrutability of databases along with the sometimes rigid storage demands and preferences of database administrators often serve to stymie these efforts. It can seem as if a kind of Mars/Venus chasm exists between the two groups.

Recently, Oracle Corp. introduced a rather unique tool called the Oracle ILM Assistant that may help to bridge this gap. First of all, one feature that should make it immediately of interest is cost -- there is none. However, it also represents the first product that I've come across that tries to pragmatically address the real challenge of value-based mapping of database contents to storage.

The traditional approach to database-storage alignment has typically been to establish generic classification policies, e.g. production databases reside on Tier 1 storage, development and test on Tier 2, etc. Oracle ILM Assistant (download PDF), on the other hand, offers a straightforward graphical user interface to more granularly partition databases and assign each partition to an appropriate storage tier.

It also contains a couple of additional features that are likely to be attractive to both the storage manager and the database administrator. The storage manager can assign a per-gigabyte cost for each proposed tier of storage, and the tool can then model and report the storage cost savings of tiered partitioning based on specific database tables and classification policies. For the risk-averse administrator, the tool doesn't actually move data, but instead generates the necessary SQL scripts that could be executed to partition the database (assuming it is not already partitioned). The database administrator can examine, test and become comfortable with these scripts before executing thereby adding an important level of comfort.

Certainly, this tool isn't appropriate for all applications, but given the growing number of database-driven applications, by following the logical process of defining tiers, establishing life-cycle policies, selecting tables and generating scripts, it can be useful for improving database storage efficiency. And perhaps also help to mend a troubled database administrator/storage manager relationship.

Jim Damoulakis is chief technology officer of GlassHouse Technologies Inc., a leading provider of independent storage services. He can be reached at jimd@glasshouse.com.

Copyright © 2007 IDG Communications, Inc.

7 inconvenient truths about the hybrid work trend
Shop Tech Products at Amazon