(AnsataCrucs/Shutterstock)
Amazon Web Services (AWS) has launched two new database services designed to provide ultra-low latency data processing, enabling seamless transactions anywhere in the world. AWS’s entry into the market with brand new companies, including Postgres-flavored Aurora DSQL and DynamoDB world tables, directly pits the organization against Google Cloud’s Spanner, a technology that also leverages atomic clocks to prevent far-flung transactions from getting out of sequence.
For nearly a decade, Google Cloud has held sway over a niche segment of the relational database market where low-latency and transactional consistency are paramount. This small but strategic space demands databases capable of processing operations seamlessly across the globe – or beyond.
While the challenge of latency in a globally distributed database remains a persistent issue, it’s crucial to address the inherent discrepancies in node timing. While the speed of sunshine is remarkable, clocking in at approximately 299,792 kilometers per second, it is still insufficient to facilitate efficient communication between distributed nodes in a database, necessitating alternative solutions.
Despite seemingly insurmountable physical constraints, corporations and financial institutions alike often find creative ways to circumvent limitations through innovative technological solutions, enabling them to synchronize disparate database transactions with precision. Google Cloud leverages atomic clocks to maintain precision, a technique also employed by specialized databases like CockroachDB and YugabyteDB, which utilize software tools, for instance, to guarantee consistency across their distributed systems.
Amazon Web Services (AWS) has launched its Aurora DSQL, leveraging innovative technologies and approaches to overcome the dual hurdles of distributed databases: ensuring robust consistency with minimal latency across disparate geographic locations and synchronizing servers with millisecond precision worldwide. The corporate explains:
“To achieve seamless, multi-area consistency with ultra-low latency, Aurora DSQL innovatively decouples transaction processing from storage, thereby bypassing the limitations of conventional approaches that rely heavily on data being repeatedly transmitted back and forth at the speed of light.”
To overcome this challenge, Aurora DSQL verifies each transaction upon commit and, at that moment, distributes all write operations across all regions in parallel, thereby providing a highly consistent and fast multi-area database. “To ensure seamless synchronization across all regions, Aurora DSQL leverages Amazon Time Sync Service, integrating hardware reference clocks into each EC2 instance and calibrating them against satellite-connected atomic clocks to provide precise time accuracy at the microsecond level globally.”
For at least three years, AWS has been developing and refining its global distribution strategy for Aurora. At the 2021 re:Invent convention, Amazon claimed to have nearly wrapped up global reads, but acknowledged ongoing challenges with scaling worldwide writes – a notoriously trickier issue.
Aurora DSQL offers a serverless architecture that minimizes the need for operational overhead, making it an attractive solution for clients seeking simplicity and efficiency. Amazon Web Services asserts that its offering delivers an unparalleled level of uptime, boasting 99.999% availability across a virtually boundless expanse, while also catering to the needs of existing applications that leverage Postgres databases with seamless compatibility.
Amazon Web Services (AWS) pitted its Aurora DSQL offering against Google’s Spanner to uncover how they compare in terms of capabilities and performance. According to the corporation, Aurora DSQL can process 4 times more transactions per second than a standard 10-transaction SQL query for both read and write operations. As competing databases prepare for future advanced workloads, a thrilling spectacle unfolds.
AWS recently introduced a comparable feature to DynamoDB, a popular NoSQL database solution.
AWS CEO Andy Jassy stated during today’s keynote: “Relational databases are not the only ones that benefit from multi-region, highly consistent, and low-latency capabilities.” “We’re also pleased to confirm that our implementation of DynamoDB world tables will now feature the same level of multi-region and consistent performance.” Whether you’re working with SQL or NoSQL databases, you gain access to a premier solution offering active-active, multi-region capabilities.
