AWS has released an updated Aurora Serverless platform that speeds capacity scaling by 45 percent and boosts database performance by up to 30 percent, delivering cost benefits through improved scheduling and smaller ACU increments.
AWS has rolled out an updated Aurora Serverless platform that speeds scaling and lifts throughput. The change arrives as cloud teams fine‑tune cost models and response times for variable workloads.
The new platform version 4 delivers about 45 percent faster capacity ramp‑up when demand spikes and raises database performance by up to 30 percent. AWS cites improved runtime efficiency and a smarter scaling algorithm that makes capacity adjustments in smaller steps of 0.5 ACUs. The update applies automatically to new clusters and can be enabled for existing clusters through the ServerlessV2PlatformVersion parameter.
A HammerDB TPROC‑C benchmark measured NOPM across three platform versions with 1,024 virtual users. Results show 27‑34 percent higher NOPM for Aurora MySQL and Aurora PostgreSQL on version 4 compared with version 3. A Sysbench test loaded 250 tables totaling 16 GB and ran 50 million read‑heavy queries with 512 threads. The study reported 27 percent faster completion and 28 percent lower cost versus version 3, and 41 percent faster completion with 42 percent lower cost versus version 2.
Corey Quinn, chief cloud economist at The Duckbill Group, noted in his newsletter that Aurora Serverless now scales up 45 percent faster and can scale down to zero capacity, which aligns with his view of serverless databases that bill in ACU fractions. He added that the improvements come at no extra charge, suggesting competitive pressure or an accidental PRFAQ approval.
Pini Dibask, principal database solutions architect at AWS, highlighted on LinkedIn that Database Savings Plans announced at re:Invent 2025 provide up to 35 percent discount for Aurora Serverless, combining with the performance gains of Platform Version 4 to shift the cost equation.
Renato Losio, editor at InfoQ and recognized AWS Data Hero, wrote the article. He has worked as a cloud architect and tech lead in Berlin and contributes regularly on cloud services and relational databases.
The update reflects a broader trend of cloud providers refining serverless database offerings to meet demand for automatic scaling and cost efficiency. For teams building web applications and API services, the faster ramp‑up and higher throughput can reduce latency during traffic bursts and lower ACU consumption when workloads are light.
You can read the original AWS blog post here: AWS blog post
More details on the HammerDB benchmark are available at: HammerDB benchmark
The Sysbench results can be found here: Sysbench results
Comments
Please log in or register to join the discussion