RESEARCH

RESOURCES / RESEARCH

Helping Different Companies With Information Management

McKnight Consulting Group Report: High-Concurrency TPC-C Benchmarking

November 10, 2026

Today’s servers can pack 96 or more processor cores into a single chip. On hardware that big, the question isn’t whether the machine is powerful enough. It’s whether the database can actually use that power. A database that gets in its own way leaves expensive capacity sitting idle, and you pay for computing power you never get to use.

In this benchmark report from McKnight Consulting Group, an independent research firm, sponsored by MariaDB, MariaDB Enterprise Server 12.3 went head to head with its predecessor, MariaDB Enterprise Server 11.8, and with MySQL Community Edition 9.7.2. All three ran on the same large cloud server under heavy load, and the results are clear:

56% more work at its best than MariaDB 11.8. At its best result, MariaDB 12.3 handled 56% more orders per minute than the previous version.

About 25% ahead of MySQL at the same load. With the same number of simultaneous users, MariaDB 12.3 handled about 25% more orders per minute than MySQL 9.7.2 between 64 and 120 users, and 21% more at 152 users. At each database’s best result, MariaDB’s lead was 21%.

Get the report here.

The TCO and ROI of GPU infrastructure

November 10, 2026

The True Engineering Cost of Open Model Training: A TCO and ROI Analysis of GPU Infrastructure compares Crusoe Cloud, a competing neocloud, and a hyperscaler on cost, effort and reliability across the full cluster lifecycle.

The rate card tells you what an hour of compute costs. It doesn’t tell you how many engineers it takes to run the platform, or how much of that hour turns into finished training.

So the engineers ran a common test suite on all three clouds, from a cold-account deployment to a 48-hour, 24-GPU training run on Llama 3.1 70B, with failures injected along the way. Then they modeled what a lean team would spend over three years.

Get the report here.

iPaaS: Integration Platform-as-a-Service Functionality Evaluation

November 10, 2026

An integration platform-as-a-service (iPaaS) product is a cloud-based platform that enables organizations to quickly and easily integrate data across different systems or applications and automate corresponding business processes.

We evaluated two leading iPaaS solutions, Boomi and Workato, across the top 29 tasks a business would perform with the tools. For our ratings, we considered functionality, speed of development, and ease of use. Our tests of the most frequently used iPaaS workflows revealed that Workato produced a 29% advantage over Boomi in functionality, development, and user experience.

Get the report here.

Agentic Data Management at Scale: Benchmarking Vector and Hybrid Search Across Postgres and Specialized Databases

September 19, 2026

McKnight Consulting Group benchmarks put EDB Postgres AI ahead of dedicated vector databases, lakehouse platforms, and every other tested Postgres provider on speed, accuracy, and cost for production agents.

Get the report here.

Measuring Data Fabric TCO: How a Unified Business Data Fabric Provides the Foundation for AI 

April 2, 2026

The modern data landscape is currently characterized by a “great unbundling,” where architectural flexibility often comes at the cost of seamless integration. This fragmentation frequently leads to “pilot purgatory,” where data initiatives stall due to the technical friction of treating data as a byproduct rather than a high-velocity capital asset. 

In this benchmark, we compared two paths to deliver the same end-state outcome:

  • SAP Business Data Cloud (SAP path): Implemented in a provisioned tenant using a defined finance scenario, capturing elapsed time, engineering effort, and platform consumption aligned to published pricing. 
  • DIY data fabric (DIY path): Architected to production standards and decomposed into build-and-run tasks across ingestion, modeling, governance, semantics/graph, analytics, application enablement, and agent orchestration. Effort and cost were modeled using published service pricing and standardized burdened labor assumptions.

Contact us for the report.

Independent  Benchmark Report: Impetus LeapLogic™ Outperforms Native Migration Tools 

March 3, 2026

Leads in conversion accuracy, effort reduction, and end-to-end timeline:
A study by McKnight Consulting Group. hows unmatched real-world advantages over native migration tools in head-to-head comparison, saving up to 52 weeks of conversion time for large migrations.

Get the report here.

Benchmark: Aerospike vs. ScyllaDB

February 5, 2026

Aerospike was provisioned with approximately one-third fewer infrastructure resources than ScyllaDB in all tests, due primarily to differences in recommended replication factors. Despite this, Aerospike consistently outperformed ScyllaDB across throughput, latency, and stability metrics, resulting in substantially better performance per unit of provisioned capacity.

Get the report here.

Aerospike vs. DynamoDB: See the benchmark results

January 26, 2026

DynamoDB performance struggles to keep up at scale, and its pricing only gets worse as you grow. If your applications demand predictable low latency, high throughput, and operational affordability, Aerospike is the better choice.

This independent benchmark by McKnight Consulting Group compares Aerospike, DynamoDB, and DynamoDB with DAX across 1 TB, 5 TB, and 10 TB datasets for both 70/30 and 100/0 read/write workloads. The results are clear: Aerospike outperforms on every front, including latency, throughput, and total cost of ownership, at every scale.

Get the report here.

Why DIY MDM Approaches Stall — and What the Data Says About a Better Path in Fabric

October 2, 2025

DIY MDM can work for a single domain or POC, but it stalls at scale with governance gaps, mounting tech debt and slower delivery.

This independent McKnight benchmark report shows Profisee for Microsoft Fabric cuts 3-year TCO by up to 55% (44% for large projects) and delivers in ~12 weeks vs. 9-52 weeks for DIY.

The report maps four real-world Fabric scenarios and pinpoints when to move beyond DIY to accelerate time-to-value with built-in governance.

Get the report here.

Performance and Cost Benchmarking of Real-Time Data Platforms 

October 2, 2025

Microsoft Fabric Real-Time Intelligence Compared with a Competitive Build Solution Using Confluent Kafka and Flink Running on AWS.

Contact us for the report.