A Decade in Review: Distributed SQL Takes the Stage as NewSQL Exits

A Decade in Review: Distributed SQL Takes the Stage as NewSQL Exits

Few players remain standing 10 years after the term NewSQL was first used to describe a new generation of relational database management players that came to market to challenge the likes of Oracle, IBM and Microsoft.

451 Research’s Matt Aslett, Research Director of data, AI and analytics, looks back at how NewSQL fared, and the rise of the distributed SQL vendors.

NewSQL vs. NoSQL

NewSQL was introduced in early 2011 to refer to a group of emerging database products and vendors that could be further grouped into three sub-categories: novel systems built from the ground up from new architecture, middleware that reimplements the same sharding infrastructure developed by Google and others, and distributed database-as-a-service offerings from cloud providers.

In comparison, NoSQL is more purpose-built to support requirements for flexible application development and data processing that were unsuited to the relational database model. In a way, NewSQL was intended to combine the scalability advantages of NoSQL with the structure, consistency, performance and transactional support provided by the relational data model.

In 2016, Aslett worked with Andy Pavlo, associate professor of  Databaseology in the Computer Science Department at Carnegie Mellon University to develop the definitive descriptions of NewSQL, including architecture trends that drove the NewSQL databases, products and services, and how NewSQL differentiates from the traditional relational databases.

Few survivors

In hindsight, pioneers of NewSQL were ahead of customers’ demand for distributed data processing and database managed services.

VoltDB and SingleStore remain the handful of NewSQL vendors that survived out of the many that Aslett mentioned in a 2011 report. VoltDB now operates within a narrower focus on telecommunications and 5G use cases, while SingleStore continues to focus on hybrid workloads for both OLTP (online transactional processing) and OLAP (online analytical processing).

Others have undergone a rebranding. Database sharding specialist CodeFuture came to be known as  AgilData in 2015, and turned into a cloud- and data-centric services provider. JustOne became Edge Intelligence in 2017 to focus on analytics at the edge.

Going global

Newcomers known as distributed SQL subsequently entered the market with a rearchitected concept of relational databases for a globally distributed architecture. Taking a leaf from the more successful NoSQL vendors, the distributed SQL vendors deepened their engagement with customers to educate them about the user-friendly features and benefits of their products. Here are the descriptions of the notable players:

  • Cockroach Labs: Founded by former Google engineers, the company offers CockroachDB for self-hosted deployment on cloud infrastructure and CockroachCloud managed services. Baidu, Bose, Comcast, DoorDash, eBay and SpaceX are its customers.
  • PlanetScale: A developer-focused database service.
  • PingCAP: Launched the public preview of its TiDB Cloud managed services.
  • YugaByte: Open source services are now available in EMEA and APAC.
  • Fauna: Managed serverless database is available via an SQL API.

In addition, the horizontally scalable cloud database services offered by the cloud giants including AWS Amazon Aurora, Google Cloud Spanner and Azure Database for PostgreSQL-Hyperscale are also considered distributed SQL.

Seize the demand

This second generation of SQL is likely to succeed where its NewSQL predecessors have failed. A survey by the 451 Alliance found that 58% of IT decision-makers expected to use a globally distributed database within a year, while 70% planned to do so within three years. It is time to move on.