PGConf.SPB 2023

Fuzzing research is feeding random input data to a program (or a part of it) (in fact, randomness is very conditional) and seeing what we get out of it. And we repeat it many times on many processors.

Fuzzing a large monolithic program complex is never a simple task. It requires extraordinary solutions. In this talk, I will tell you what and how we searched with the help of fuzzing and what results it led to.

• Investigation of data type parsing functions (input-functions): for warming up;
• Investigation of functions implementing operations between types (op-functions): it is better to consider the structure here;
• Network subsystem fuzzing: let's pretend we are POSIX calls, it's cheaper that way;
• Recovering disk context: we need Groundhog Day.

A story about funny bugs and ridiculous hand gestures will be included.

As you probably know, PostgreSQL has a number of distinct features compared to other DBMSs. For example, Postgres can process and store many different types of data. However, you need to know something about them before using them. In this talk, we will find the reason why queries to the table begin to slow down (and autovacuum / vacuum has nothing to do with it) and try to speed up such queries. I will tell you how integer data types work in PostgreSQL and touch on the topic of speeding up such queries. And finally, let's talk about how to make your data in tables take up less space while increasing the speed of queries to this data.

In modern data analysis, machine learning models are used as often as databases. Such IT giants as Google and Amazon have already combined them. Microsoft and Yandex are not far behind. Isn't it time to implement a machine learning model in PostgreSQL? In the report you will hear about the basics of machine learning, its implementation in databases and an example of realization as postres extension.

Unfortunately, ideal computer systems exist only in science fiction books. PostgreSQL is not exception and sometimes problems may occur. I would like to discuss how to correctly (and incorrectly) try to solve a problem, which way to choose, which tool to use.

The talk is of interest to both beginners and experienced users and database administrators.