Surya Hebbar Email & Phone Number

India

Bengaluru, Karnataka, India

Bengaluru, Karnataka, India

Apache Impala is a low latency and high concurrency query engine, deployed on hundreds of nodes, that scales linearly even in multi-tenant environments, used for BI/analytics queries. It was introduced as an equivalent of Google's planet level distributed database Google F1/Spanner(used in Google Ads).• Developed graphical plan exports in SVG for large query plans that can contain more than 1 million elements in the expression tree.• Developed impala daemon's CPU utilization time series graph for all impala nodes.• Rewrote the entire fragment timing diagram in SVG, reforming the canvas based code, after greatly improving the rendering time by decoupling the retrieval and rendering of fragments.• Devised an algorithm to align the CPU utilization time series samples with the fragment timing diagram.• Supported scaling precision of time ticks and utilization values for the entire query timeline, while maintaining visual and rendering constraints.• Refactored the code to render all of the above components parallely and asynchronously for better performance.• Supported scaling of all of the above components horizontally and vertically with mouse and keyboard events.• Supported SVG exports of the entire query timeline consisting of the fragment timing diagram and the aligned cpu utilization chart.• Manual testing of all of these on TPC-DS and TPC-H queries