The Next Wave of SQL
Technology cycles map over five to ten year horizons, especially in the enterprise where group product decisions last years and even old ideas need time to resurface.
We’ve seen this recently as the hype of Hadoop has long passed, and even NoSQL has hit certain limits.
Today, tried and trusted approaches with SQL are in vogue, and a wave of new startups, from early to later stage, are bringing this universal lingua franca for data and analytics back into the fold.
Last week I took a brief look at three companies (Cockroach Labs, YugaByte, and Crate.io) focusing on Distributed SQL, an approach to transactional workloads in global environments. In this post, I’ll share a new set of companies employing SQL differently across their solutions.
First, two companies focusing on database and analytics solutions
Timescale bills itself as “Simple, scalable SQL for time-series and IoT” and touts “The power of SQL for time-series data.”
Like many startups in this group, Timescale built on top of proven technologies instead of starting from scratch. Specifically they make use of PostgreSQL, one of the world’s most popular open-source databases, and a solid choice for datasets that benefit from robust SQL analytics.
Rockset describes its product as “Serverless search and analytics for event data,” and “Millisecond-latency SQL on operational data from Kafka, DynamoDB, S3 and more.”
Rockset provides a layer between data sources on one side (streams, lakes, databases, data warehouses) and data consumers on the other (applications, dashboards, notebooks) using RocksDB, “a high performance embedded storage engine optimized for SSDs. RocksDB is used in production at Facebook, LinkedIn, Yahoo, Netflix, Airbnb, Pinterest and Uber.”
Next two companies addressing the ETL (Extract, Transform, Load) market
Segment identifies “Customer Data Infrastructure” as an umbrella to describe the myriad of tools and data sources companies employ today to get a complete customer picture. Segment provides an abstraction that speeds this process and ideally gets companies a more complete view of customer data quickly, or in their words, “to simplify the mess of analytics APIs.”
Segment also recently introduced a new feature they call SQL Traits, “a new feature in Personas that allows you to use SQL to pull customer data directly from your warehouse into Personas and activate it in your marketing tools.”
Fivetran helps customers “build robust, automated pipelines with standardized schemas that free you to focus on analytics, not ETL.” And they do so allowing you to “access to all your data in SQL.”
Similar to Segment, Fivetran has a number of pre-built connectors to simplify the process of getting data from different data sources into a customer’s data warehouse.
Shifting to data modeling, analytics, and the end-to-end data lifecycle
Sigma Computing is a cloud business intelligence and analytics tool that allows you to “Connect to your warehouse in minutes and start analyzing billions of rows of data — including JSON — using the full power of SQL and the cloud.”
Sigma also includes a data modeling capability to “Build centralized data definitions and curated data views.” As companies look to integrate data from a wide variety of sources, developing consistent models at enterprise scale will remain a perennial challenge and opportunity.
Moving to the application layer we have another interesting SQL-focused entrant
Transposit touts itself as, “The first API composition platform, empowering developers to easily build apps in the expanding API ecosystem.” There have been API centric solutions before, but Transposit is taking a novel approach by bringing SQL to the API layer.
SQL the SQL
SQL dates back to the early 1970s and the famed paper by researcher E.F. Codd from IBM. Part of the reason SQL has lasted this long is how simply it represents many mathematical concepts in English. The math hasn’t changed much over the years, meaning that SQL, in a range of new forms, is here to stay.