Percona Live Online

Open source software is the defacto standard for many new applications, this is especially true in the database industry. Currently, MySQL, PostgreSQL, MariaDB, MongoDB, Elastic, and others have shown up in every industry and organization in the world in some form or another. People are no longer choosing a single database for the company, they are letting developers and architects choose the best database for the job.

This has led to an increase in the number of technologies operations teams have to support. Couple that increases in technologies with a growing micro-service ( or cloud-native ) development paradigm where every service has its own database and where all the data is valuable.

Now companies are now faced with dozens of technologies, hundreds or even thousands of individual database instances, and petabytes of data. The management of the complexity of such an environment is changing the way we look at systems and operations.

Let’s talk about the trends and tell you what you need to know about how to manage the new multi-verse of data.

Web technologies have come leaps and bounds. But are you still using the tired old database from the last generation? Let's look at the methodology of microservices, compare it to bounded contexts, and look at ops tasks for micro-databases. Let's tour all the flavors of databases, understand their pros and cons, and when you would choose it. You'll leave with a roadmap for moving from data-monolith to micro-databases.

Why store when you can stream? Modern open source streaming platforms like Apache Kafka provide means of storing and processing event driven data. When combined with streaming analytics projects like Apache Flink, many business applications, and event driven microservices may not even need a database storage layer. With the rise of streaming SQL engines, in particular FlinkSQL, developers with SQL and database skills can use these to build complex event processing and advanced analytics applications. There are performance advantages to being able to run queries before data lands in a database. However, full solutions often need long term storage as well for regulatory purposes, forensic purposes or to train machine learning to work better in the stream. This talk will cover new approaches to data architecture, how it ts in to existing event and traditional database applications, and provides practical examples, and approaches to operating systems built on streaming on cloud, using open source

For a considerable set of applications querying geographical data consists of a critical operation. Fast responses combined with a high level of accuracy are often the requirements when an application user interacts with functions/operations of the type “Give me near me” or “Find me in area XYZ”. Additional complexity is usually added when the points of interest are constantly on the move, like a public transportation vehicle or a taxi.

For applications that frequently access geographical data and rely on both speed and accuracy, both application and database design is crucial. In this presentation, we are going to focus on the database side. More specifically, we are going to evaluate three of the most popular open-source databases, MongoDB, Postgres, and Elastic against geospatial workloads. For each of these databases, we are going to examine the implementation and the performance of geo-queries. We are going to discuss best practices and design patterns for each database and try to find a winner among the three.

OLAP and OLTP databases provide different trade-offs for database users, but what if you could have the best of both worlds?

In this talk we describe how to use ClickHouse as an OLAP replication slave of MySQL databases, including support for transactional consistency (MVCC- which ClickHouse does not support). The talk will go into how the MaterializeMySQL database engine in ClickHouse works, the modifications we did to support transactional consistency, and the options users will have for choosing between freshness and OLAP performance.

ARM is gaining a lot of traction, especially with High-Performance Computing Softwares.

Opensource Databases is no exception and most of the leading opensource databases are now available on ARM (MySQL, MariaDB, PostgreSQL, MongoDB, ClickHouse, etc...)

Let's explore the state of different open-source databases and their supporting ecosystems/tools, understanding the performance, functionality, active community, etc...

Whatever your use-case it is quite likely that it could be ported to ARM and this comes with a lot of advantages.

So let's unwind this completely new VERTICAL of running Opensource DBs on ARM.

During this talk we will discuss how the index works on Postgres and MySQL. What are the differences between the implementations and what are the more appropriate choices for different scenarios? We will discuss the general B+-tree indexes but also discuss GIN, GIST and understand where they are best suited with examples and understand why some database migrations are a failure due to differences in implementation or lack of a specific index.