Platform

Predictive Analytics

• Designed and implemented target clustering based on metrics, providing the ability to quickly analyze a large collection of targets to determine which ones have similar characteristics.

• Seasonality aware anomaly detection: improved alerting by considering seasonal highs and lows, added to REST API based analytical service.

• Project lead for integrating multiple components (data collection, platform, analytics, UI) for database stress index prediction.

Oracle GoldenGate, XStream, and Streams

• Automatic parallelism tuning: enabled dynamic optimization for scaling the number of apply processes based on throughput.

• Project lead for procedural replication and large raw/varchar values.

• Designed and implemented support for conflict detection and resolution for inserts, updates and deletes for multi-master configurations for distributed databases.

• Implemented transaction level exception handling for OGG.

Streams (distributed database replication and information sharing)

• Implemented eager apply of changes at the destination database to reduce latency for large transactions – providing seconds latency instead of minutes for transactions with 1 million or more changes.

• Project lead for replication for XMLType, which is also used as the basis for object replication.

• Improved performance by caching apply process metadata, used for processing all row changes.

• Prevented excess memory consumption, and minimized restart time by spilling messages for large and long running transactions.