| TPC-DS V4: Embracing the Dynamic Pricing Model (DPM) |
The cornerstone of TPC-DS V4 is the integration of the Dynamic Pricing Model (DPM) . In the cloud era, a DPM represents a fundamental departure from traditional fixed-capacity provisioning, moving instead toward a consumption-based "pay-for-what-you-use" architecture. By decoupling compute resources—such as CPU and RAM—from storage, this model allows each to scale independently. Unlike legacy systems that require manual intervention or reboots to scale, dynamic systems monitor real-time workload metrics to adjust hardware allocation automatically. This eliminates the "idle tax"—the unnecessary cost of maintaining peak-load hardware during low-activity periods—and aligns infrastructure expenditure directly with business value. |
Impact on Performance Evaluation |
|
For industry-standard benchmarks like TPC-DS, which simulate complex decision-support systems and big data environments, dynamic pricing introduces a transformative variable into performance evaluation. Historically, TPC-DS results were reported as a static Price/Performance metric, where the fixed cost of a hardware configuration was divided by query throughput. Under a DPM, the "Price" component becomes fluid. A system can now leverage massive resources to accelerate the intensive Power Test (complex joins and aggregations) while scaling down to near zero during the Maintenance Test (data refreshes). This shift prioritizes Total Cost of Ownership (TCO) modeling, rewarding databases that demonstrate "elastic efficiency"—applying maximum power only when the workload demands it and relinquishing it the moment a task concludes. |
Key Specification Changes |
|
Supporting a DPM required structural revisions to both the Pricing Specification and the TPC-DS Specification:
|
|
Specification
|
|
More Information
|
|
Results
|





