Guest: Mike Freedman (LinkedIn)
Company: Timescale (Twitter)
Show: Let’s Talk

Timescale is a modern cloud platform built on PostgreSQL for time series, events, and analytics. In this episode of TFiR: Let’s Talk, Timescale Co-Founder and CTO Mike Freedman shares his insights on the current database landscape and how Timescale enables developers to build faster and scale efficiently.

Current data trends in the market:

• Modern applications are data intensive. A lot of ingestion is happening, with events from the product, usage metrics from devices, sensors, and other real-time data feeds. Applications are then built to use all these data to provide either live analytics or live data that’s available for query.
• The amount of data being ingested has nothing to do with the size of the company. A 10-person startup could be dealing with tens or hundreds of terabytes, all in the interest of serving this type of modern applications.

On differentiating databases:

• When people talk about databases, they often refer to OLTP, OLAP, and HTAP, which are meant to describe the workloads. Freedman believes that’s not the right way to think about it. Instead, databases should be differentiated in terms of who will use it and for what purpose.
• Is the database going to be used by the developers or the data team? There is a big distinction.
• Developer teams who are building applications on top of a database and that database has to be operationalized and productized. They need faster queries, dashboards, and flexibility to build the best solution for the application that they’re developing.
• Data teams deal with centralized data infrastructure using traditional data warehousing, data lakes, or lake houses. They bring together all these disparate sources from across the organization, some product data, some sales data across ETL jobs to have a centralized view for reporting, analysis, and model training.

Timescale is Postgres++:

• Timescale is Postgres that is rearchitected for modern applications in modern cloud, i.e., it supports everything Postgres does and adds new capabilities such as native compression, hybrid row columnar storage, vectorized execution, and continuous aggregates.
• It is unique compared to other companies playing in this space because it has teams that extend the internals of the database as well as teams that engineer around the database in the infrastructure itself.
• Its client base is a combination of major enterprises to small startups, including HP, Warner, Uber, Coinbase, Northvolt, and Lucid.
• It supports financial and crypto, manufacturing and IoT, transport and logistics, gaming, music security, and observability use cases.

Backstory:

• Timescale did not start out as a database company focusing on time series. They were actually building a platform for IoT sensor data. Back then, they were looking at some of the open-source solutions that provided time-series databases.
• Most companies in that space deployed both a time-series-only database and a relational database like Postgres. Whenever they want to ask new questions about their data, they need to deploy a new application code that pulls data from two different databases and then joins them.
• They built TimescaleDB on Postgres to store both their time-series events and analytical data, as well as their business and metric data. It gave them the SQL, the performance, and the operational simplicity they wanted.

How Timescale helps developers:

• The maturity that it has through Postgres provides a breadth of features and functionality that is powerful for companies, not just when they start, but as they scale their business and their applications. It leads to engineering productivity and helps reduce business risks.
• It achieves 90-95% compression because it uses algorithms designed for the specific types of data, i.e., it compresses different types of data differently.
• It takes SQL queries and some of that get executed in main memory on standard disk, while others get transparently pushed down into S3.
• It has support and ops in APAC, EMEA, and North America for follow-the-sun models of all of its cloud infrastructure.

This summary was written by Camille Gregory.