Citus BlogCitus Blog

Around 10 years ago I joined Amazon Web Services and that’s where I first saw the importance of trade-offs in distributed systems. In university I had already learned about the trade-offs between consistency and availability (the CAP theorem), but in practice the spectrum goes a lot deeper than that. Any design decision may involve trade-offs between latency, concurrency, scalability, durability, maintainability, functionality, operational simplicity, and other aspects of the system—and those trade-offs have meaningful impact on the features and user experience of the application, and even on the effectiveness of the business itself.

Perhaps the most challenging problem in distributed systems, in which the need for trade-offs is most apparent, is building a distributed database. When applications began to require databases that could scale across many servers, database developers began to make extreme trade-offs. In order to achieve scalability over many nodes, distributed key-value stores (NoSQL) put aside the rich feature set offered by the traditional relational database management systems (RDBMS), including SQL, joins, foreign keys, and ACID guarantees. Since everyone wants scalability, it would only be a matter of time before the RDBMS would disappear, right? Actually, relational databases have continued to dominate the database landscape. And here's why:

Postgres has been a great database for decades now, and has really come into its own in the last ten years. Databases more broadly have also gotten their own set of attention as well. First we had NoSQL which started more on document databases and key/value stores, then there was NewSQL which expanding things to distributed, graph databases, and all of these models from documents to distributed to relational were not mutually exclusive. Postgres itself went from simply a relational database (which already had geospatial capabilities) to a multi modal database by adding support for JSONB.

But to me the most exciting part about Postgres isn't how it continues to advance itself, rather it is how Postgres has shifted itself from simply a relational database to more of a data platform. The largest driver for this shift to being a data platform is Postgres extensions. Postgres extensions in simplified terms are lower level APIs that exist within Postgres that allow to change or extend it's functionality. These extension hooks allow Postgres to be adapted for new use cases without requiring upstream changes to the core database. This is a win in two ways:

1. The Postgres core can continue to move at a very safe and stable pace, ensuring a solid foundation and not risking your data.
2. Extensions themselves can move quickly to explore new areas without the same review process or release cycle allowing them to be agile in how they evolve.

Okay, plug-ins and frameworks aren't new when it comes to software, what is so great about extensions for Postgres? Well they may not be new to software, but they're not new to Postgres either. Postgres has had extensions as long as I can remember. In Postgres 9.1 we saw a new sytax to make it easy to CREATE EXTENSION and since that time the ecosystem around them has grown. We have a full directory of extensions at PGXN. Older forks such as which were based on older versions are actively working on catching up to a modern release to presumably become a pure extension. By being a pure extension you're able to stay current with Postgres versions without heavy rebasing for each new release. Now the things you can do with extensions is as powerful as ever, so much so that Citus' distributed database support is built on top of this extension framework.

This year, as part of the Citus Data annual team retreat and hackathon, we decided to get ambitious and try to create an easy, 1-click migration to the Citus distributed database. My team was lucky enough to get Dimitri Fontaine, the author of the excellent PostgreSQL pgloader utility, so we decided to start our hackathon project by building on top of pgloader and making pgloader more Citus-aware.

For those of you not familiar with Citus—Citus is an extension to PostgreSQL that transforms Postgres into a distributed database. We make Citus available as open source (star us on GitHub!), as enterprise software you can run anywhere, and as a fully-managed database as a service.

Materialized views were a long awaited feature within Postgres for a number of years. They finally arrived in Postgres 9.3, though at the time were limited. In Postgres 9.3 when you refreshed materialized views it would hold a lock on the table while they were being refreshed. If your workload was extremely business hours based this could work, but if you were powering something to end-users this was a deal breaker. In Postgres 9.4 we saw Postgres achieve the ability to refresh materialized views concurrently. With this we now have fully baked materialized view support, but even still we've seen they may not always be the right approach.

Today, we’re excited to announce that we have donated 1% of Citus Data’s stock to the non-profit PostgreSQL organizations in the US and Europe. The United States PostgreSQL Association (PgUS) has received this stock grant. PgUS will work with PostgreSQL Europe to support the growth, education, and future innovation of Postgres both in the US and Europe.

To our knowledge, this is the first time a company has donated 1% of its equity to support the mission of an open source foundation.

To coincide with this donation, we’re also joining the Pledge 1% movement, alongside well-known technology organizations such as Atlassian, Twilio, Box, and more.

At Citus whether it's looking at our own data or helping a customer debug a query I end up writing a lot of SQL. When I do write SQL I do my best to make sure it's readable in case others need to come along and understand or modify, but admittedly I do have some bad habits from time to time such as using implicit joins. Regardless of my bad habits I still try to make my SQL and database as easy to understand for someone not already familiar with it. One of the biggest tools for that is comments.

Even early on in learning to program we take advantage of comments to explain and describe what our code is doing, even in times when it seems obvious. I see this less commonly in SQL and databases, which is a shame because data is just as valuable so making it easier to reason and work with seems logical. Postgres has a few great mechanisms you can start leveraging when it comes to commenting so you can better document things.

Before joining the Citus Data team to work on the Postgres extension that transforms Postgres into a distributed database, I checked out the Citus Data blog to learn a bit more. And found all sorts of material for Postgres beginners as well as power users, with one of my favorites being the "Fun with SQL" series. After I joined, I jumped at the chance to write a Fun with SQL post, and the only question was what to write about. I chose unions.

In SQL, the UNION and UNION ALL operators help take multiple tables and combine the results into a single table of all matching columns. This operator is extremely useful if you want the results to come back as a single set of records.

Our beloved Structured Query Language may be the lingua franca for relational databases—but like many languages, SQL is in a state of constant evolution. The first releases of SQL didn't even have a notation for joins. At the time, SQL only supported inner joins.

Cross Joins and Where Filters

As a result, back in early eighties, the only way to express a join condition between tables would be in the WHERE clause.

Citus extends Postgres to be a horizontally scalable database. By horizontally scalable, we mean the data is spread across multiple machines, and you're able to scale not only storage but also memory and compute—thus providing better performance. Without using something like Citus to transform PostgreSQL into a distributed database, sure you can add read replicas to scale, but you're still maintaining a single copy of your data. When you run into scaling issues with your Postgres database, adding a read replica and offloading some of your traffic to your read replica is a common bandaid to slow down the bleeding, but it is only a matter of time until even that doesn't work any further. Whereas with Citus, scaling out your database is as simple as dragging a slider and rebalancing your data.

Are read replicas still useful with horizontally scalable databases?

But that leaves a question, are read-replicas still useful? Well, sure they are.

PostgreSQL 11 is brewing and will be released soon. In the meantime, testing it with your own application is a great way to make sure the community catches all the remaining bugs before the dot-zero release.

One of the big changes in the next PostgreSQL release is the result of Andres Freund's work on the query executor engine. Andres has been working on this part of the system for a while now, and in the next release we are going to see a new component in the execution engine: a JIT expression compiler!