We are pleased to announce a significant new release of Simora, version 3.0. This release brings just one feature, but it’s a huge one: The ability to playback a captured workload (from an Oracle database) against a non-Oracle test database. In this initial release we are focused on supporting the Impala engine, with all testing taking place with Impala 2.0.1 running on a Cloudera CDH5 distribution. Other databases will be added as minor releases since Simora can now support any ODBC-compliant database.
A number of new features have been implemented to support heterogenous playback:
- Support for ODBC datasources
- SQL distillation: All SQL is collated into a single source file to allow one-stop editing of SQL syntax and efficient caching of SQL
- SQL parser support: SQL is automatically rewritten to support features of the target datasource (where practical)
- Integration of the Simora user interface with the Cloudera Manager metrics API
This release is focused on the replay of Data Warehouse workloads, answering the exam question:
Will an Imapala/Hadoop cluster support the execution of my entire Data Warehouse workload?
By capturing an entire day of Data Warehouse traffic and replaying with Simora, it is now possible to know if the workload can be supported by the cluster.
Simora 3.0 is currently in development, and available on early release in conjunction with Scale Abilities consulting services.