Performance

What performance should I expect?

Up to your tier's bandwidth ceiling, with consistency that comes from the node being dedicated to you. Real throughput depends on distance to the node, object sizes, and how many transfers you run at once.

Why is a single download slower than my tier?

One stream rarely saturates a high-bandwidth link because of latency and per-connection limits. Running several transfers in parallel is the single biggest lever for getting close to your ceiling.

How do I get the most out of my tier?

• Run parallel transfers (e.g. rclone --transfers, AWS CLI s3 sync).
• Choose the closest region to reduce latency.
• Prefer fetching larger objects over many tiny files (less per-request overhead).
• Test from a machine with enough bandwidth of its own.

See Throughput & speed for step-by-step tuning.

Does distance to the region matter?

Yes. Higher latency to a far-away region reduces real throughput per connection, even on a large tier. Pick the nearest region.

Will lots of small files be slower than a few big ones?

Generally yes — each object has request overhead, so thousands of tiny files spend more time on round-trips than on data transfer. Batching or using parallel transfers helps.

Is my performance affected by other customers?

No — your node and its bandwidth allocation are dedicated, so you're not sharing throughput. That's the main reason performance stays predictable.

How can I benchmark my node?

Run a parallel transfer of representative data and measure aggregate throughput, not a single-stream copy. A single cp understates what your tier can do.

I think my node is underperforming — what now?

Confirm region, concurrency, and your own local bandwidth first. If it still looks off, email support@clearstreamer.com with details and they can check your dedicated node.