Deep Dive into AWS Aurora DSQL

3rd December, 2024, AWS announced the preview of Amazon Aurora DSQL. It is still not generally available.

In today’s hyper-connected world, businesses need global scalability, low latency, and high availability to serve distributed user bases seamlessly. AWS Aurora DSQL (Distributed SQL) is a groundbreaking evolution of AWS Aurora, designed to address the challenges of modern global applications.

With Aurora DSQL, AWS now provides an alternative to its customers for Multi-Region Proprietary Relational Database product category which had only 1 player till now, Cloud Spanner from Google.

Here’s a deep dive into Aurora DSQL :

What is AWS Aurora DSQL?

Aurora DSQL builds on the core capabilities of Amazon Aurora, introducing a distributed architecture that supports global write scalability and low-latency reads and writes across multiple regions. With multi-master support, Aurora DSQL enables active-active configurations, ensuring consistency, availability, and scalability for the most demanding workloads.

How Aurora DSQL Achieves Global Scalability

Aurora DSQL is designed to handle massive workloads with a distributed architecture optimized for scalability, performance, and availability:

1. Distributed Storage Layer: Aurora DSQL’s shared storage spans multiple Availability Zones (AZs) and regions, allowing seamless replication and petabyte-scale data management.
2. Multi-Master Capability: Supports concurrent writes across multiple regions, enabling up to hundreds of thousands of write requests per second. Multi-master clusters allow active-active writes, ensuring low-latency performance for global applications.
3. Multi-Region Clusters: Handles up to 15 read replicas per region and can replicate data across regions with latencies as low as 10-20 milliseconds. Supports millions of read requests per second, making it ideal for read-heavy applications with distributed users.
4. Horizontal Scaling: Aurora DSQL can dynamically scale compute and storage resources independently. Compute scaling adds nodes to handle increasing workloads, while storage scales automatically up to 128 TB per database cluster.
5. Global Transaction Coordination: Ensures ACID compliance for distributed writes using a consensus protocol, maintaining strong consistency across all regions. Aurora DSQL can process tens of thousands of transactions per second (TPS) in global-scale scenarios.
6. Dynamic Workload Balancing: Intelligent routing ensures optimal utilization of nodes, balancing workloads across regions and replicas.
7. Fault Tolerance and Recovery: Provides sub-second failover for read replicas and under 30-second failover for writer nodes, ensuring 99.99% availability.

How is Aurora DSQL Different from AWS Aurora?

Applications Benefiting from Aurora DSQL

1. Global-Scale Web and Mobile Apps : Social networks, content delivery platforms, and collaboration tools.
2. Data-Intensive Systems : Applications requiring distributed analytics or large-scale reporting.
3. Business-Critical Services : Systems that demand fault tolerance and consistent, reliable data access.
4. Cross-Border Compliance : Applications needing data residency across multiple countries or regions.

Choosing the Right Solution

Choose Aurora When:

• You need a cost-efficient, high-performance database for single-region applications.
• Your workloads are primarily read-heavy within a single geographical area.

Choose Aurora DSQL When:

• You require global scalability with low-latency access for distributed users.
• Applications need active-active writes and fault tolerance across regions.
• High availability and multi-region compliance are business-critical.

Aurora DSQL extends the capabilities of Aurora, making it a powerful solution for modern, global, and high-demand applications.