US 2024/0256531 A1Published application

Execution and attestation of user defined functions in databases

AssigneeDatabricks, Inc.

InventorsMartin Grund, Herman Rudolf Petrus Catharina van Hövell tot Westerflier, Stefania Leone

Life of this patent

Estimating the term…

Loading patent…

Abstract

A system executes user defined functions (UDFs) invoked by database queries. The UDF includes UDF code specified using a programing language distinct from a database query language. A hash value from the UDF code provided by a client application for creating the UDF is compared with a hash value generated from UDF code invoked by database queries to determine whether the two UDF codes match. If the two hash values fail to match, the system takes an action, for example, storing an indication of UDF code mismatch or disabling subsequent executions of the database queries invoking the UDF. The system may use encoded UDF code that is decoded by the system at runtime using a key obtained from a separate system such as the client application. The client application can disable execution of database queries executing the UDF code by refusing to provide the key.

Claims20

Dependent claims

A computer-implemented method for executing user defined functions in database systems, the computer-implemented method comprising:

receiving, from a client application, a request to create a user defined function, the user defined function (UDF) comprising a first UDF code specified using a programing language distinct from a database query language, wherein the UDF is configured to be invoked using database queries specified using the database query language;
generating a first hash value based on the first UDF code, the first hash value generated using a hash function and an encryption key;
receiving, by a database system, a request to execute a database query that invokes the UDF;
determining a second UDF code executed during execution of the database query;
generating a second hash value based on the second UDF code executed during execution of the database query, the second hash value generated using the hash function and the encryption key;
comparing the first hash value with the second hash value to determine whether the first UDF code specified while creating the UDF matches the second UDF code executed during execution of the database query;
and
responsive to the first UDF code specified while creating the UDF failing to match the second UDF code executed during execution of the database query, storing an indication of a UDF code mismatch.

5 more elements

A non-transitory computer readable storage medium comprising stored instructions for executing user defined functions in database systems that when executed by one or more computer processors cause the one or more computer processors to:

receive, from a client application, a request to create a user defined function (UDF) comprising a first UDF code specified using a programing language distinct from a database query language, the UDF is configured for invocation using database queries using the database query language;
generate a first hash value based on the first UDF code, the first hash value based on a hash function and an encryption key;
receive, by a database system, a request to execute a database query invoking the UDF;
determine a second UDF code executed during execution of the database query;
generate a second hash value based on the second UDF code executed during execution of the database query, the second hash value based on the hash function and the encryption key;
compare the first hash value with the second hash value to determine whether the first UDF code specified upon generation of the UDF matches the second UDF code executed upon execution of the database query;
and
store an indication of a UDF code mismatch when the first UDF code specified upon generation of the UDF fails to match the second UDF code upon execution of the database query.

5 more elements

A computer system comprising:

a computer processor;
and
a non-transitory computer readable storage medium comprising stored instructions for executing user defined functions in database systems that when executed by one or more computer processors cause the one or more computer processors to:
receive, from a client application, a request to create a user defined function, the user defined function (UDF) comprising a first UDF code specified using a programing language distinct from a database query language, the UDF is configured for invocation using database queries using the database query language;
generate a first hash value based on the first UDF code, the first hash value based on a hash function and an encryption key;
receive, by a database system, a request to execute a database query invoking the UDF;
determine a second UDF code executed during execution of the database query;
generate a second hash value based on the second UDF code executed during execution of the database query, the second hash value based on the hash function and the encryption key;
compare the first hash value with the second hash value to determine whether the first UDF code specified upon generation of the UDF matches the second UDF code executed upon execution of the database query;
and
store an indication of a UDF code mismatch when the first UDF code specified upon generation of the UDF fails to match the second UDF code upon execution of the database query.

8 more elements

Dependencies are inferred from the claim text.