@inproceedings{karanasos2020extending,
author = {Karanasos, Konstantinos and Interlandi, Matteo and Xin, Doris and Psallidas, Fotis and Sen, Rathijit and Park, Kwanghyun and Popivanov, Ivan and Nakandal, Supun and Krishnan, Subru and Weimer, Markus and Yu, Yuan and Ramakrishnan, Raghu and Curino, Carlo},
title = {Extending Relational Query Processing with ML Inference},
booktitle = {Conference on Innovative Data Systems Research (CIDR)},
year = {2020},
month = {January},
abstract = {The broadening adoption of machine learning in the enterprise is increasing the pressure for strict governance and cost-effective performance, in particular for the common and consequential steps of model storage and inference.

The RDBMS provides a natural starting point, given its mature infrastructure for fast data access and processing, along with support for enterprise features, such as encryption, auditing, and high-availability. To take advantage of all of the above, we need to address a key concern: Can in-RDBMS scoring of ML models match (outperform?) the performance of dedicated frameworks?

We answer the above positively by building Raven, a system that leverages native integration of ML runtimes (such as ONNX Runtime) deep within SQL Server and a unified intermediate representation (IR) to enable advanced cross-optimizations between ML and database operators. In this optimization space, we discover the most exciting research opportunities that combine DB/compiler/ML thinking. Our initial evaluation on real data demonstrates performance gains of up to 5.5× from the native integration of ML in SQL Server and up to 24× from cross-optimizations. An early preview of the ONNX Runtime integration is currently available with Azure’s SQL Database Edge.},
url = {http://approjects.co.za/?big=en-us/research/publication/extending-relational-query-processing-with-ml-inference/},
}