Froid header: database methods balanced on a scale

Froid

Froid is an extensible, language-agnostic framework for optimizing imperative functions in databases. Currently, the benefits of UDFs (User-Defined Functions) such as modularity, readability, code reuse come with a huge performance penalty. The purpose of Froid is to enable developers to use the abstraction of UDFs without compromising on performance. Froid introduces a radically different approach to evaluating imperative UDFs that results in improvements of up to multiple orders of magnitude over the current state of the art.

Aggify (opens in new tab) is a related project that broadens the scope of Froid and expands its applicability by allowing optimization of cursor loops in UDFs.

Froid is available as «Scalar UDF Inlining» in Microsoft SQL Server 2019 (Read the documentation) (opens in new tab)

 

Talk by Karthik at IIIT Bangalore (Oct 2018) as part of RMIT 2018

Talk by Karthik at UW Seattle (Mar 2019) as part of the NWDS Database Lecture Series (opens in new tab)

 

Talk by Prof. Andrew Pavlo (CMU) as part of the CMU Advanced Database Systems course: Spring 2019 (opens in new tab), Spring 2020 (opens in new tab)