At AWS re:Invent 2024, the global cloud leader AWS unveiled a series of innovative upgrades to its cloud database products, aiming to enhance functionality while reducing enterprise operational costs. These innovations include the new Amazon Aurora DSQL distributed SQL database, the global table feature for Amazon DynamoDB NoSQL database, and multi-region capabilities for Amazon MemoryDB. AWS also elaborated on its overall database strategy and explored the pivotal role of vector databases in advancing generative AI applications. Additionally, AWS announced a series of pricing adjustments, including up to a 50% reduction in on-demand pricing for Amazon DynamoDB.
AWS's innovations are not just designed to attract the attention of database administrators but, more importantly, to address real-world application challenges. These new features are part of AWS's comprehensive strategy to tackle the growing and complex workloads in distributed deployments. AWS's cloud database portfolio particularly emphasizes support for workloads with extremely high real-time requirements. During the keynote speech, multiple AWS users, including United Airlines, BMW, and the NFL, shared how they utilize AWS cloud databases to enhance business efficiency.
Ganapathy (G2) Krishnamoorthy, AWS Vice President of Database Services, stated during the conference, "We are committed to simplifying database management through innovation, allowing you to focus on building the next generation of applications. Databases are a critical component of applications and a key part of our vision for data analytics and artificial intelligence."
In the distributed SQL sector, AWS is redefining how this architecture operates to accelerate read and write operations for always-available applications. Unlike traditional distributed databases, Aurora DSQL employs a leader-less architecture, achieving unlimited scalability. This innovation is based on the Firecracker micro-VM technology that powers AWS Lambda's serverless capabilities, allowing each system component to scale independently.
AWS also addresses consistency issues in distributed databases through an approach known as "optimistic concurrency." In this method, all database operations run locally, and only transaction commits occur across regions, thereby preventing interference with the entire application due to excessive log retention. Krishnamoorthy noted, "When designing Aurora DSQL, we built optimistic concurrency into its core, eliminating the need for locks to ensure consistency or isolation."
Additionally, AWS has introduced strong consistency and global distribution capabilities to the DynamoDB NoSQL database. DynamoDB global tables with strong consistency synchronize write data to persistently store in multiple regions, enabling applications to read the latest data from any region. This feature allows mission-critical applications to deploy across multiple regions without requiring changes to application code.
United Airlines is one of the loyal users of AWS's new features. In a video testimony at the conference, Sanjay Nayar, the company's Managing Director, stated that they use AWS to manage over 2,500 database clusters, store more than 15PB of data, and run millions of transactions per second. United Airlines leverages DynamoDB global tables as an essential component of its passenger seating system to ensure fast and reliable writes and reads of seat allocation information.
Amazon MemoryDB in-memory database has also been upgraded with added multi-region capabilities. AWS offers vector support in a range of cloud databases, but Amazon MemoryDB stands out in performance. This is why the NFL chose this database to help build AI-based generative applications. Eric Peters, NFL's Director of Media Management and Post Production, stated that they use MemoryDB for short-term memory queries and save successful queries to vector storage for future use. These stored memories can guide new queries, leading to faster and more accurate results.
AWS's innovations not only enhance the functionality and performance of cloud databases but also offer greater value to enterprise users by reducing prices.
