Exercise 3: Compare Quanton vs OSS Spark

See Quanton's performance acceleration in action by running all 99 TPC-DS queries (1 GB scale factor by default) against both engines on the same cluster, then comparing per-query wall-clock time. The benchmark suite is included in the Quanton Operator repo and prints a speedup table at the end.

# Build the datagen image and load it into minikube (one-time)
docker build -t tpcds-datagen:latest -f benchmarks/Dockerfile.datagen benchmarks/
minikube image load tpcds-datagen:latest
minikube image load apache/spark:3.5.0

# Run — defaults to scale factor 1 (~5 min total)
./benchmarks/run.sh

# Or use a meatier scale factor for a more realistic comparison (~30-45 min)
./benchmarks/run.sh --scale-factor 10

The script runs four phases sequentially: data generation → OSS Spark queries → Quanton queries → results table.

Watch the Spark UI as it runs

While a phase is in flight, port-forward to the live driver pod:

kubectl get pods -A | grep driver

# During Phase 3 — OSS Spark queries
kubectl port-forward oss-spark-tpcds-driver 4040:4040

# During Phase 4 — Quanton queries
kubectl port-forward quanton-tpcds-parquet-driver 4040:4040

Then open http://localhost:4040. Re-bind as the script transitions phases.