PgConf.Russia 2020

Corosync & pacemaker is a well known solution for creating fault-tolerant clusters. Such clusters can contain 3 working nodes or 2 working nodes and one voting-only node. The cluster can be deployed on physical or virtual servers.

This tutorial will demonstrate the process of installation and tuning of a PostgreSQL fault-tolerant cluster. You will learn that it is not so difficult as seems to be from the first glance.

Role-based access control (RBAC) is one of the main mechanisms used for access control in many DBMS, including PostgreSQL. This model is a sub-type of traditional discretionary access control with its restrictions. In addition to DAC, many operating systems also use mandatory access control (MAC) based on security labels. This additional security mechanism is obligatory for protecting information that demands higher levels of security. Naturally, we would like to use MAC within DBMS when working in OS with mandatory access control switched on.
In this talk, we'll give an overview of existing MAC implementations in DBMS, as well as share our approach to using security mechanisms provided by SELinux, the sepgsql extension for PostgreSQL, and the standard mechanism of row-level security (RLS), which has been available in PostgreSQL starting from version 9.5.
In our presentation, we will use the "Airlines" demo database provided by the Postgres Professional company to show how to protect sensitive information and personal data, compare different ways of storing security labels, and assess performance of our solution.

If you care about Postgres performance, there are a number of hardware acceleration options to help with different use cases. Intel Optane DC persistent memory creates new tier in data hierarchy allowing developers to utilize performance of traditional memory combining with volume and persistency of block storage devices. Unlike traditional DRAM-only in-memory systems, where memory is small, expensive, and volatile, Intel Optane DC persistent memory makes it possible to run larger Postgres databases (terabytes) in memory for higher performance. FPGAs are integrated circuits that can be reprogrammed dynamically to accelerate a specific workload such as SQL execution and data compression. FPGA accelerators extend Postgres with hundreds of SQL reader and writer processes that work in parallel on the FPGA. It’s similar to adding hundreds new cores to boost parallel processing on your server.

Recently, I was working on a project where graphQL was used for sending requests to its .NET Core backend, but this was not a good idea. The point is, a graphQL query is a hierarchical structure with a dynamic set of fields. It’s difficult to perform such requests via a statically-typed programming language and a relational database as suggested by the tools available. So, I came up with the idea of using the plv8 extension and perform graphQL queries right on the database side. It took me about two hours to develop a working prototype that could perform the same queries as the software under development for more than one month! Then various improvements have been made and I want to introduce them all. If you are thinking of using graphQL instead of REST, my speech could be most useful and could help you to save a lot of time.