Introduction

Quanton is a storage-aware execution engine for Apache Spark, built by Onehouse. It combines native vectorized execution with optimizations that use Iceberg and Hudi metadata, indexes, and table layouts to reduce the data your jobs read, shuffle, and rewrite.

Your Scala, PySpark, and Spark SQL jobs run unchanged. You run Quanton on your own infrastructure, with your existing data, catalogs, and orchestration.

Why storage awareness matters

A lakehouse job spends time fetching data, executing transformations, and writing results. Quanton optimizes all three types of work:

• Read: prefetch data within the read path and decode table data in native columnar batches.
• Transform: reshape query plans to remove redundant work and use available indexes to find matching rows.
• Write: reduce shuffle during MERGE and combine base data with updates through native columnar compaction.

For example, an indexed merge can locate the files and row groups containing matching keys before scanning the target. A low-shuffle merge then reduces the data moved across the network to apply those updates. These optimizations work alongside vectorized operators.

Read Engine architecture for the mechanisms and examples, or Comparisons to evaluate the alternatives.

How it fits into your stack

Quanton runs on Kubernetes, including EKS, GKE, AKS, and local minikube, and can run inside an existing Spark platform. The Quanton Operator integrates with the Spark Operator's SparkApplication workflow to configure jobs for the Quanton runtime. The operator manages deployment integration; the engine executes your jobs.

The AI Spark engineer is available on every plan, including the free tier, with your own model API key. It uses job context in the Spark UI to diagnose bottlenecks and recommend tuning changes.

Pricing is per GiB processed or a flat annual fee. See pricing and the current benchmark results.

Get started

Path	When to use
Local Quickstart	Try Quanton on your laptop with minikube
EKS Guide	Deploy on Amazon Elastic Kubernetes Service
GKE Guide	Deploy on Google Kubernetes Engine
AKS Guide	Deploy on Azure Kubernetes Service
Benchmarking	Run the TPC-DS suite on your own cluster