Why SQL Is Making a Comeback: From NoSQL’s Rise to the New Data Era

After years of seeming dormancy, SQL is experiencing a revival that could reshape the future of data engineering and analytics.

Since the advent of cloud computing, the exponential growth of data highlighted SQL’s inability to scale horizontally, sparking the NoSQL movement with technologies such as MapReduce, Bigtable, Cassandra, and MongoDB.

Despite the hype, signs now indicate that SQL is returning. Major cloud providers offer managed relational databases—Amazon RDS, Google Cloud SQL, Azure, and Alibaba Cloud RDS. Amazon’s Aurora, compatible with PostgreSQL and MySQL, is among the fastest‑growing AWS services, while SQL interfaces on Hadoop, Spark, and Kafka continue to thrive.

New Hope

Understanding SQL’s comeback begins with its origins in the early 1970s at IBM Research, where Donald Chamberlin and Raymond Boyce created a user‑friendly query language that later became SQL in 1974.

SQL quickly dominated with systems like System R, Ingres, DB2, Oracle, SQL Server, PostgreSQL, and MySQL, becoming the de‑facto language for interacting with databases.

NoSQL’s Counterattack

While Chamberlin and Boyce were developing SQL, engineers in California were building ARPANET, the precursor to the modern Internet. The rapid expansion of the web in the late 1980s generated massive, fast‑growing data streams that relational databases struggled to handle.

This led to the creation of distributed, non‑relational systems: Google’s MapReduce (2004) and Bigtable (2006), Amazon’s Dynamo (2007), followed by Hadoop, Cassandra, and MongoDB. These NoSQL solutions promised scalability but introduced new challenges: proprietary query languages, lack of mature features like JOIN, increased glue code, and fragile ecosystems.

Back to SQL, Leaving the Dark Temptation

SQL resurfaced through interfaces on Hadoop and Spark, promoting the “Not Only SQL” (NoSQL) concept. The NewSQL wave followed, with horizontally scalable OLTP databases such as H‑Store (2008), Google Spanner (2012), CockroachDB (2014), and extensions to PostgreSQL (JSON support, partitioning, parallelism, etc.).

Projects like TimescaleDB demonstrate how extending PostgreSQL can meet modern data‑intensive workloads, reinforcing SQL’s relevance.

Google’s Example

Google’s Spanner evolved from a NoSQL‑style API to a full‑featured SQL system, sharing a standard SQL language across internal services like F1, Dremel, and external tools such as BigQuery, thereby reducing friction for developers and analysts.

SQL as a Universal Interface

Just as IP serves as a universal networking interface, SQL is emerging as the universal data interface, enabling diverse components of the technology stack—from storage engines to analytics tools—to interoperate with minimal overhead.

In today’s data‑rich world, developers face fragile glue code and high maintenance costs when stitching together heterogeneous systems. Embracing SQL reduces this complexity and provides a readable, widely understood language for data manipulation.

Conclusion

SQL is undeniably back, not merely because NoSQL’s shortcomings are painful, but because the world is awash with data that demands a common, powerful, and accessible query language.

We should embrace both NoSQL and SQL, leveraging each where it shines while recognizing SQL’s role as the backbone of modern data infrastructure.