What’s Next for Databases? Exploring Innovative Platforms for 2024

Relational database management systems have dominated for decades, but 2023 introduced disruptive, often distributed, databases that blend SQL, NoSQL, and serverless concepts. The article reviews several innovative platforms and compares their characteristics.

PlanetScale

PlanetScale is a serverless MySQL platform built on Vitess, Google’s horizontally scalable open‑source database. It introduces a "branch" workflow similar to Git, allowing developers to create isolated database branches for experiments and merge them back once changes are verified. Key features include automatic daily backups, non‑blocking schema changes, and seamless integration with Prisma. Pricing starts with a free tier (1 production and 1 development branch, 1 billion reads and 1 million writes per month) and scales to paid plans ($29–$599 per month). Limitations include missing MySQL features like stored procedures and a learning curve for distributed concepts.

YugaByteDB

YugaByteDB is an open‑source, PostgreSQL‑compatible database designed for cloud‑native environments, supporting multi‑cloud deployments to avoid vendor lock‑in. It offers both SQL and NoSQL APIs, high performance, scalability, and strong availability. Managed clusters cost $0.25 per core‑hour.

Dolt

Dolt treats databases like Git repositories, enabling version‑controlled data with branching, merging, and standard Git commands (e.g., dolt log, dolt add). It supports SQL queries and collaborative editing, automatically merging changes from multiple developers. Dolt is open‑source and free, but lacks advanced features such as stored procedures.

CockroachDB

Developed by former Google engineers, CockroachDB is a distributed SQL database compatible with PostgreSQL, built in Go. It supports multi‑model APIs, multi‑cloud deployment, and can handle millions of queries per second. A free tier exists for small projects; the dedicated plan costs $295 per month. Drawbacks include operational complexity, a steep learning curve, and sub‑optimal performance for ultra‑low‑latency workloads.

Cloudflare D1

Cloudflare D1 is a serverless, edge‑run SQLite‑compatible database that stores data as document‑like objects. It offers a key‑value store for Workers and allows JavaScript‑based stored procedures. Currently in alpha, D1 is free for testing but not production‑ready, with limited query capabilities and migration challenges.

Xata

Xata combines PostgreSQL and Elasticsearch, providing full‑text search, a visual schema editor, and branch support similar to Dolt. It targets real‑time, high‑volume workloads. Pricing includes a free tier (75 requests/second, up to 750 k records) and a professional plan at $8 per unit per month. Migration difficulty and overall complexity are noted drawbacks.

SurrealDB

Written in Rust, SurrealDB offers a unified API for relational, document, and graph data. It excels in flexible data modeling for recommendation engines and social networks, and provides a SQL‑like query language (SurrealQL). It is free for self‑hosting, though cloud hosting is pending and community support is limited.

FaunaDB

FaunaDB is a serverless NoSQL database with native relational capabilities, its own query language (FQL), and built‑in GraphQL support. It uses a distributed architecture for high availability and low latency. A free tier offers 100 k reads, 50 k writes, and 5 GB storage; paid plans range from $25 to $500 per month. The proprietary query language introduces a learning curve.

Conclusion

These databases largely adopt distributed architectures, requiring additional setup and expertise, but they provide innovative workflows such as branching, serverless operation, and multi‑model support. Choice depends on pricing, ease of use, community backing, and specific use‑case needs: document‑oriented workloads may favor FaunaDB, branching‑centric development can use PlanetScale or Dolt, and flexible schemas point to Xata.