Speed & Secure Sybase Operations
Optimize ETL & Reorgs, Mask PII, Replicate Data
Challenges
As an SAP Sybase IQ or ASE DBA, you may face one or more of these performance or protection issues:
- Unloading and loading of large Sybase tables
- Slow utility operations (e.g., reorgs) or queries
- Unprotected personally identifiable information (PII)
- Cumbersome database migration or replication
- Generation of, or access to, test data sets
Specific performance diagnoses and tuning also take time, and may affect other users. Specialized data masking and test data management tools are expensive and too hard to use. Finally, stored SQL procedures may also be programmed inefficiently, require optimization, and then still take too long to run.
Solutions
- Speed Sybase Unloads
- Speed Transforms, Queries, and Loads
- Speed Sybase Reorgs
- Migrate & Replicate Sybase Data
- Mask PII in Sybase
- Generate & Populate Sybase Test Data
IRI FACT (Fast Extract) to dump transaction tables in parallel to flat files. FACT is many times faster than other extract methods and is supported in the common IRI Workbench front-end for standalone use, or with other data-driven IRI operations like ETL, reorg, and migration.
Use the SortCL program in the IRI CoSort product or IRI Voracity platform to filter, sort, join, and aggregate the extracts in parallel. The Sybase SQL Server Reference Manual recommends external pre-sort to accelerate index creation during database loads. Multiple tables created quickly in query index order should help speed queries.
How to Speed Sybase Loads: "CoSorting" data in the file system prior to loading Sybase can:
- Speed load performance 2 and 10X
- Remove the sorting aspect of the load
- Reduce database and disk overhead
- Increase load processing rate (rows per hour)
The recommended technique says to pre-sort flat files on the clustered index key. CoSort is designed to sort huge files in parallel, with performance scaling linearly in volume. You will want to create a clustered index. But first, eliminate the indexes by creating tables without indexes or triggers, or by dropping them from the table to be loaded. Then you can:
- CoSort the data on the primary key index
- bcp load the sorted table using fast bcp path (as there are no indexes on teh table)
- Use the SQL command CREATE INDEX, using the SORTED_DATA option, to create the clustered index
CoSort's SortCL program can also transform and reformat data for database migrations and custom (detail, summary, and delta) reports, and protect sensitive data at the field level with a variety of encryption, masking, and other functions.
Accelerate reorgs by performing them offline; i.e., use the file system instead of encumbering the database. The offline DB reorg wizard available to FACT and CoSort (or the platform Voracity) users in the IRI Workbench combines FACT, CoSort and bcp in batch operations you can schedule and run in or outside the Workbench GUI.
Move data in and out of Sybase and other data sources with IRI NextForm DB edition or the platform Voracity. Perform complex ETL, change data capture, and replication operations in Voracity.
Find and classify PII through built-in data discovery and profiling tools. Then, build static data masking jobs for Sybase, SAP HANA, and other JDBC and ODBC-connected sources with the standalone IRI FieldShield package or Voracity platform. You can also score the risk of re-identification based on the remaining quasi-identifiers in your dataset.
For dynamic data masking, make a call to FieldShield functions in its SDK.
To rapidly populate Sybase with safe test data, use the high-volume test data generator, IRI RowGen. RowGen uses your data models to automatically generate the test data for an entire database with referential integrity, and is also built into the Voracity platform.
Blog Links
Other Resources