DBI Launches New Version of Brother-Owl

DBI (Database-Brothers, Inc.), a software company providing professional grade database performance management tools, has announced the availability of release 7.1 of Brother-Owl-a tool to capture, analyze, and fix Oracle application SQL and indexes.

Release 7.1 adds support for Oracle and 11g as well as HP-UX 11i Itanium. Many new dynamically computed performance metrics were also added to aid the DBA in rapid, accurate, isolation of costly SQL. Beyond the new platform support, support for Linux is also expected to be available this summer, notes Scott Hayes, president and CEO of DBI, noting that Linux on x86 platforms is planned. "We are also going to work on z/Linux," he adds.

Comparing a database workload to a grocery cart which is filled with one item that is very expensive, but whose price is exceeded by the aggregate cost of many units of an inexpensive  item,  Hayes states,  "The art of what DBI does  is  we help the customer accurately  identify and isolate that work in a database which costs the most." According to Hayes, "What we uniquely do is help DBA teams accurately understand the cost in the database workload." Equipped with the knowledge of what costs the most in the SQL statement shopping cart, "DBA teams can then make effective decisions about what statements truly require their tuning attention," he explains.

"The analysis that we do is really very illuminating. Once we have helped the DBA identify the most costly statement, then we provide them with a very simple to use, integrated workflow where we can do explains, we can generate hints, we can generate index advice and we can do benchmarking of different alternatives-all with out any impact to the production database. It is all done in a simulated manner."

Additionally, states Hayes, the next most important benefit DBI brings to the table is Brother-Owl's ability to read Oracle's memory directly. "The tool is lightning-fast and we collect the performance information from Oracle without any performance burden or overhead on the database." To learn more, go here.