Blog
Shubham Sinha, Co-Founder

Benchmarking SQL Server CDC to ClickHouse

Up to 7x faster than Fivetran and Airbyte.

Change Data Capture (CDC) from SQL Server to ClickHouse is one of the most common replication patterns we see at Supermetal. As more customers evaluate Supermetal for this workload, we benchmarked both snapshot and CDC at scale against Fivetran and Airbyte.


Overview

Supermetal is deployed as a single Rust binary. It uses Arrow-native primitives for performance and object storage for durability. No need to wrangle Kafka, Debezium, or other JVM-based tooling. Read more in the architecture docs.

SQL ServerSQL Server Source
  • Native CDC: Reads directly from SQL Server's internal change tables.

  • Parallel snapshots: Automatically split large tables on the primary key and read them concurrently across all available cores.

  • Broad edition support: SQL Server 2016+, including AWS RDS, Azure SQL Database, Azure SQL Managed Instance, and standard on-prem editions.

View SQL Server docs →

ClickHouseClickHouse Target
  • Native deduplication: Updates and deletes are handled natively using ClickHouse's ReplacingMergeTree engine. No manual merge logic or row-level mutations.

  • Zero-copy writes: Supermetal stages Parquet files on object storage (S3), and ClickHouse pulls directly. No extra network hop through the replication process.

View ClickHouse docs →


Benchmark Setup

We used the TPC-H dataset at scale factors 20 through 50 (31 GB to 77 GB) to measure performance across two scenarios:

SQL ServerSQL Server
RDS db.m5.2xlarge
8 vCPU, 32 GB
10K IOPS (io2)
us-west-2
SupermetalSupermetal
EC2 m8azn.xlarge
4 vCPU, 16 GB
us-west-2
ClickHouseClickHouse
ClickHouse Cloud
32 GiB × 3 replicas
us-west-2
SQL ServerSQL Server
RDS db.m5.2xlarge · 8 vCPU, 32 GB · 10K IOPS (io2) · us-west-2
SupermetalSupermetal
EC2 m8azn.xlarge · 4 vCPU, 16 GB · us-west-2
ClickHouseClickHouse
ClickHouse Cloud · 32 GiB × 3 replicas · us-west-2

Snapshot Performance

We ran a full snapshot (backfill) at scale factors 20 through 50 from SQL Server to ClickHouse Cloud, and tested Supermetal against Fivetran and Airbyte at each scale factor.

Sync duration by vendor

Supermetal completes every tested scale factor 4–7× faster.

SF 20 · 173M rows
Supermetal4m 33s
Airbyte23m 46s
Fivetran25m 10s
SF 30 · 260M rows
Supermetal6m 38s
Airbyte32m 38s
Fivetran37m 11s
SF 40 · 346M rows
Supermetal9m 56s
Airbyte44m 07s
Fivetran53m 44s
SF 50 · 433M rows
Supermetal11m 44s
Airbyte50m 54s
Fivetran76m 28s

The gap widens at every scale factor: up to 7x faster than Fivetran and 5x faster than Airbyte.

SQL Server to ClickHouse snapshot throughput

Source reads and ClickHouse writes during the initial snapshot.

SQL Server to ClickHouse snapshot throughput0/s4.2M/s8.4M/s12.7M/s16.9M/s04:03:2804:15:07
Rows inRows out
1.3M/s
Peak throughput

CDC Performance

We tested CDC across sustained rates from 1K to 30K ops/sec. Fivetran and Airbyte use batched scheduling (typically 5–60 minute sync intervals), making their real-time CDC latency not directly comparable.

SQL Server to ClickHouse CDC throughput

Target load compared with rows processed during CDC.

SQL Server to ClickHouse CDC throughput0/s16K/s31K/s47K/s62K/s06:21:0006:53:40
Target loadProcessed rows
63K/s
Peak throughput
4.5s
P95 latency
SQL Server to ClickHouse latency breakdown

End-to-end, source-read, and target-write latency during CDC.

SQL Server to ClickHouse latency breakdown0ms3.1s6.3s9.4s12.5s06:21:0006:53:40
End-to-endReadWrite
63K/s
Peak throughput
4.5s
P95 latency

Try Supermetal

Supermetal ships high-performance connectors for SQL Server and ClickHouse, deployed as a single binary on any machine. No Kafka, no JVM, no cluster. The trial includes 1,000 hours of free sync.

curl -fsSL https://trial.supermetal.io/install.sh | sh
iwr -useb https://trial.supermetal.io/install.ps1 | iex

Quickstart · SQL Server docs · ClickHouse docs · Architecture